Five Fatty Aldehyde Dehydrogenase Enzymes from Marinobacter and Acinetobacter spp. and Structural Insights into the Aldehyde Binding Pocket

ABSTRACT

Enzymes involved in lipid biosynthesis and metabolism play an important role in energy conversion and storage and in the function of structural components such as cell membranes. The fatty aldehyde dehydrogenase (FAldDH) plays a central function in the metabolism of lipid intermediates, oxidizing fatty aldehydes to the corresponding fatty acid and competing with pathways that would further reduce the fatty aldehydes to fatty alcohols or require the fatty aldehydes to produce alkanes. In this report, the genes for four putative FAldDH enzymes from Marinobacter aquaeolei VT8 and an additional enzyme from Acinetobacter baylyi were heterologously expressed in Escherichia coli and shown to display FAldDH activity. Five enzymes (Maqu_0438, Maqu_3316, Maqu_3410, Maqu_3572, and the enzyme reported under RefSeq accession no. WP_004927398) were found to act on aldehydes ranging from acetaldehyde to hexadecanal and also acted on the unsaturated long-chain palmitoleyl and oleyl aldehydes. A comparison of the specificities of these enzymes with various aldehydes is presented. Crystallization trials yielded diffraction-quality crystals of one particular FAldDH (Maqu_3316) from M. aquaeolei VT8. Crystals were independently treated with both the NAD⁺ cofactor and the aldehyde substrate decanal, revealing specific details of the likely substrate binding pocket for this class of enzymes. A likely model for how catalysis by the enzyme is accomplished is also provided.

IMPORTANCE This study provides a comparison of multiple enzymes with the ability to oxidize fatty aldehydes to fatty acids and provides a likely picture of how the fatty aldehyde and NAD⁺ are bound to the enzyme to facilitate catalysis. Based on the information obtained from this structural analysis and comparisons of specificities for the five enzymes that were characterized, correlations to the potential roles played by specific residues within the structure may be drawn.

INTRODUCTION

The fates of fatty compounds within the cell are a central aspect of lipid metabolism. Enzymes involved in the metabolism of fatty compounds can play important roles in disease and have a biotechnological relevance for both the production of lipids or biofuels and the degradation of oils that are released into the environment (1). The bacterium Marinobacter aquaeolei VT8 was isolated from the head of an offshore oil well near Vietnam, where it would be expected to be participating in the biodegradation of crude oil inadvertently released into the environment (2). In addition to oil bioremediation in the natural environment, this bacterium also produces a high-value lipid, the wax ester, which has commercial significance and is used in a range of commodity products (3). M. aquaeolei VT8 is relatively easy to culture in the laboratory, and many of the enzymes obtained from this species show strong activity when expressed in heterologous hosts such as Escherichia coli, Saccharomyces cerevisiae, and Synechococcus elongatus (4–6). For this reason, M. aquaeolei VT8 has become a model species for the study of lipid metabolism and an important source of enzymes for biotechnological applications.

Our laboratory has an interest in studying the enzymes that take part in the production of high-value compounds associated with wax ester production, including those that reduce fatty acyl coenzyme A (CoA), fatty acyl-acyl carrier protein (ACP), or fatty aldehydes to fatty alcohols (7–10). Fatty alcohols are a specific substrate for the wax ester synthase (3, 11, 12). In addition to enzymes that reduce fatty aldehydes to fatty alcohols, there are also enzymes present in M. aquaeolei VT8 that could oxidize the fatty aldehyde back to a fatty acid, such as the fatty aldehyde dehydrogenase (FAldDH). A previous report characterized a FAldDH from another species that accumulates wax esters, Acinetobacter sp. strain M-1 (13). A BLAST search using the primary sequence of the FAldDH from Acinetobacter sp. strain M-1 revealed a large number of genes from M. aquaeolei VT8 that might also function as FAldDHs (13–15). These enzymes are of particular concern in biosynthetic strategies, because they could result in a futile pathway that would diminish the accumulation wax ester or other targeted products, especially for enzymes that utilize fatty aldehydes, including fatty aldehyde decarbonylases or fatty acyl-CoA reductases and fatty aldehyde reductases (8–10, 16, 17). One particular fatty aldehyde dehydrogenase (Maqu_3410) was highly transcribed during wax ester accumulation in M. aquaeolei VT8 when grown on the simple carbohydrate citrate (CIT) or malate (8). For this reason, and because a thorough analysis of a variety of putative FAldDHs from bacteria is lacking in the literature, we sought to isolate and characterize a range of these putative FAldDHs from M. aquaeolei VT8. Furthermore, we selected one enzyme for structural studies, and here we report the crystal structures of the Maqu_3316 FAldDH from M. aquaeolei VT8 in complex with the NAD⁺ cofactor or a fatty aldehyde substrate.

RESULTS

Enzyme activity and specificity with fatty aldehydes.

The primary aim of this research was to characterize the fatty aldehyde oxidation activities of several putative FAldDH enzymes from Marinobacter aquaeolei VT8. An initial survey of the M. aquaeolei VT8 genome (15) using the BLAST algorithm (14) and the previously reported FAldDH sequence from Acinetobacter sp. strain M-1 (13) revealed a large number of genes with significant amino acid sequence identity to this previously characterized FAldDH (GenBank accession no. AB042203). Five putative FAldDH genes from M. aquaeolei VT8 and an additional gene from Acinetobacter baylyi were cloned and heterologously expressed in Escherichia coli. Each gene was modified to incorporate a polyhistidine tag (at the N terminus) to allow rapid purification. The isolated proteins obtained following metal affinity chromatography purification and a subsequent desalting step are shown in Fig. 1. A band corresponding to the expected protein size was found for each of the six proteins, although only two of the enzymes (Maqu_3316 and Maqu_3410) were obtained as high-purity single bands based on this single-step purification. Three of the enzymes (Maqu_0607, Maqu_3572, and the enzyme reported under RefSeq accession no. WP_004927398) showed minor contaminating or degradation bands, while Maqu_0438 showed the greatest degree of degradation and the lowest purity. Some variability was found for specific enzyme preparations, especially in relation to protein stability and activity. For several enzymes, the activity in certain preparations was very low, although the protein purity and quantities obtained were comparable from one preparation to another. To a certain extent, some of this activity could be rescued by degassing the protein preparation under an argon atmosphere and then adding a small quantity (5 μl) of β-mercaptoethanol to ∼8 ml of a solution containing the purified protein and allowing it to stand for several minutes. This was most prominently found for Maqu_3410, which seemed to be particularly susceptible to this issue.

FIG 1.

SDS-PAGE of various FAldDH enzymes from M. aquaeolei VT8 and A. baylyi following one-step purification by metal affinity chromatography. The protein standard is shown on the left, followed by the six enzymes, each with a primary band corresponding to a fatty aldehyde dehydrogenase. Labels include the proposed protein apparent molecular mass based on the primary sequence from the genetic construct.

An initial assessment of the substrate specificities of the M. aquaeolei VT8 enzymes using saturated, straight-chain aldehydes ranging from butanal (C₄) to hexadecanal (C₁₆) showed that four of the enzymes exhibited aldehyde dehydrogenase activity, and each of these enzymes had unique substrate selectivity patterns (Fig. 2). We also performed an alignment of these proteins with sequences of enzymes from the Protein Data Bank (PDB) that have been reported to have aldehyde dehydrogenase activity and generated a phylogenetic tree based on these alignments (Fig. 3). Maqu_0438 and Maqu_3572 clustered together in this tree based on their primary sequences and also displayed similar substrate selectivity profiles (Fig. 2) that showed a broad, almost Gaussian distribution for both proteins with straight-chain fatty aldehydes. Maqu_3316 showed a more prominent specificity for decanal, with activity dropping considerably for both octanal and dodecanal versus decanal. Maqu_3316 also clustered separately from the other enzymes tested here (Fig. 3). Maqu_3410 showed a much more consistent level of activity across the entire range, with high selectivity for longer fatty aldehydes such as hexadecanal, and was the only enzyme from M. aquaeolei VT8 that yielded a specificity profile that was similar to what was reported previously by Ishige et al. for their FAldDH enzyme from Acinetobacter sp. strain M-1 (13). In addition to the M. aquaeolei VT8 FAldDHs, we also isolated a homologous FAldDH from A. baylyi (RefSeq accession no. WP_004927398), which had very high amino acid sequence conservation with the sequence for the FAldDH enzyme from Acinetobacter sp. strain M-1. Maqu_3410 and the enzyme reported under RefSeq accession no. WP_004927398 also clustered together based on their primary sequences (Fig. 3) and displayed similar trends for their substrate selectivity profiles (Fig. 2). While our selectivity profile for the enzyme reported under RefSeq accession no. WP_004927398 increased from octanal up to tetradecanal, we also found high activity for butanal and hexanal with this enzyme and found a decrease in the activity for hexadecanal (which was not tested as a substrate for the FAldDH in that previous report [13]).

FIG 2.

Specific activities of five fatty aldehyde dehydrogenase enzymes when assayed with a range of aldehyde substrates (C₄ to C₁₆) and residues that are proposed to align the aldehyde substrate binding site. The aldehyde substrates were incrementally increased in length from left to right (x axis). Assays were run with 1.5 mM NAD⁺ and 200 μM the indicated aldehyde, and the specific activity was monitored spectrophotometrically at 340 nm in a 1-cm-path-length cuvette with a 1-ml volume. Quantities of protein added ranged from 1 μg to 100 μg. Specific activity is reported as moles of NAD⁺ reduced per minute per milligram of enzyme added (averages ± standard deviations; n ≥ 3), and reaction mixtures were maintained at 22°C. The alignment shows the regions of the various FAldDHs studied here or reported previously (13), with the residues proposed to line the aldehyde binding site highlighted. Residue numbering shown above the alignment is based on the positions for the Maqu_3316-derived enzyme. The catalytic cysteine residue is Cys282 in this enzyme.

FIG 3.

Phylogenetic tree of fatty aldehyde dehydrogenase enzymes examined in this study and six closely related aldehyde dehydrogenase enzymes found in the Protein Data Bank. The PDB accession no., GenBank accession no., or locus tag for M. aquaeolei VT8 is listed on the left, and the type of aldehyde dehydrogenase and the host organism from which it was obtained are displayed on the right. Enzymes from the Protein Data Bank with the highest similarity to those of M. aquaeolei VT8 were selected for sequence comparisons (31, 32). The primary amino acid sequences were aligned and mapped phylogenetically by using the software program Multalin (25).

Of the six genes selected for this study, all but Maqu_0607 showed FAldDH activity when a range of different aldehyde substrates was tested. Efforts to obtain activity from Maqu_0607 were abandoned after multiple purifications failed to show any activity with the substrates selected, and we instead focused our efforts on characterizing the remaining five enzymes that yielded FAldDH activity.

In addition to testing these enzymes with fatty aldehydes, we were also interested in their activity toward acetaldehyde, as many of these enzymes still showed considerable activity toward substrates as small as butanal. Of the four enzymes from M. aquaeolei VT8, all showed some residual activity toward acetaldehyde, but Maqu_3410 showed the greatest activity, which correlated with the broader enzyme activity found for this enzyme than those for the other three M. aquaeolei VT8 FAldDHs (Fig. 4). The enzyme from A. baylyi reported under RefSeq accession no. WP_004927398 showed a significantly higher level of activity with acetaldehyde than did any of the enzymes from M. aquaeolei VT8. This is in stark contrast to what was reported previously for the FAldDH from Acinetobacter sp. strain M-1, where those researchers did not report any activity with acetaldehyde (13).

FIG 4.

Specific activities of five aldehyde dehydrogenase enzymes when assayed with acetaldehyde. Assays were run with 1.5 mM NAD⁺ and 200 μM acetaldehyde, and the specific activity was monitored spectrophotometrically at 340 nm. Specific activity is reported as moles of NAD⁺ reduced per minute per milligram of enzyme added (averages ± standard deviations; n ≥ 3), and reaction mixtures were maintained at 22°C.

We were also interested in testing aldehydes that share similarity to indigenous fatty acids. M. aquaeolei VT8 contains high levels of C_16:1 and C_18:1 fatty acids (3). However, the corresponding fatty aldehydes are not easily obtained through commercial sources, so we instead pursued a strategy to synthesize these fatty aldehydes from the corresponding fatty alcohol according to an oxidation protocol.

A comparison of the specificities of each of these enzymes with unsaturated versus saturated fatty aldehydes is shown in Fig. 5. Maqu_3410, Maqu_3572, and the enzyme reported under RefSeq accession no. WP_004927398 all showed increased selectivity toward C_16:1 versus C₁₆, while Maqu_0438 and Maqu_3316 showed comparable levels of activity. All five enzymes were found to be active with C_18:1 fatty aldehyde, although Maqu_3572 was the only enzyme that showed higher activity with C_18:1 than with C₁₆. In terms of the amino acid sequence alignments, Maqu_3410 and the enzyme reported under RefSeq accession no. WP_004927398 clustered together (Fig. 3) and also showed more similar profiles with these potential native substrates containing unsaturated bonds than with the straight-chain C₁₆ fatty aldehyde (Fig. 5). Both enzymes showed the highest specific activity with acetaldehyde of the five enzymes tested here, although the activity of the enzyme reported under RefSeq accession no. WP_004927398 was significantly higher than those of all of the enzymes cloned from M. aquaeolei VT8 (Fig. 4).

FIG 5.

Comparison of substrate specificities of unsaturated aldehydes relative to the specific activities found using the saturated aldehyde hexadecanal. Assays were performed with 1.5 mM NAD⁺ and 200 μM palmityl aldehyde (C₁₆, saturated), palmitoleyl aldehyde (C_16:1, unsaturated), or oleyl aldehyde (C_18:1, unsaturated). The specific activity was measured spectrophotometrically at 340 nm, and reactions were performed at 22°C (averages ± standard deviations; n ≥ 3).

Enzyme selectivity for NAD⁺ and NADP⁺.

The oxidation of fatty aldehyde by FAldDH is dependent on the reduction of an electron-transporting cofactor. To determine if these enzymes showed any differences in selectivity toward this electron acceptor, we tested each enzyme with both NAD⁺ and NADP⁺ as electron acceptors. All five enzymes showed a higher specificity for NAD⁺ than for NADP⁺, as would be expected for catabolic processes (Fig. 6).

FIG 6.

Comparison of substrate specificities for NAD⁺ relative to NADP⁺. Each enzyme was assayed with 200 μM of the indicated aldehyde and 1.5 mM NAD⁺ or NADP⁺. The specific activity was measured spectrophotometrically at 340 nm, and reactions were performed at 22°C (averages ± standard deviations; n ≥ 3).

Optimal enzyme activity and temperature dependence.

The selectivity assays (Fig. 2 to 6) were all performed at 22°C, while a previous study (13) of a bacterial FAldDH utilized a temperature of 43°C. To determine the enzyme stability and also the maximum activity that could be obtained, assays were performed by using the substrate with the highest activity (Fig. 2), and the enzymes were further tested by ramping up the temperature until activity was no longer linear throughout the assay. Table 1 shows the relative activity at 22°C for the top substrate, the maximum activity that was obtained, and the temperature at which this maximum activity was obtained. Maqu_3572 showed the greatest temperature stability (53°C), with an ∼10-fold improvement in activity versus what was found at 22°C for decanal. A similar improvement was found for the enzyme reported under RefSeq accession no. WP_004927398 with butanal.

TABLE 1.

Activities assayed at elevated temperatures^a

Enzyme	Substrateb	Avg activity (μmol of NAD+ reduced min−1 mg enzyme−1) ± SD at 22°Cc	Avg highest activity (μmol of NAD+ reduced min−1 mg enzyme−1) ± SD (temp [°C])c
Maqu_0438	C10	0.94 ± 0.03	2.76 ± 0.18 (47)
Maqu_3316	C10	1.74 ± 0.72	3.89 ± 0.21 (43)
Maqu_3410	C6	0.70 ± 0.01	5.65 ± 0.22 (45)
Maqu_3572	C10	3.14 ± 0.13	29.25 ± 1.82 (53)
RefSeq accession no. WP_004927398	C4	3.22 ± 0.94	35.76 ± 1.93 (47)

^aReactions were performed at the highest temperature without activity loss, as described in Materials and Methods.

^bEach enzyme was assayed with an aldehyde of the indicated length, chosen for substrates with the highest specificity.

^cn = 3.

Overall structure and domain organization of the Maqu_3316 FAldDH.

Based on the purity and stability of specific FAldDH enzymes, we initiated crystallization screens and obtained two crystal structures for the Maqu_3316 FAldDH, one complexed with the substrate decanal and the other complexed with the cofactor NAD⁺. The two structures have been refined to 2.3-Å and 3.1-Å resolutions, respectively. In both structures, there are two FAldDH monomers, in the form of a symmetrical and domain-swapped homodimer, existing in an asymmetric unit. Each monomer is L-shaped and consists of 487 amino acid residues, which are organized into three domains: an N-terminal domain that binds the cofactor NAD⁺ located at the pivot of the L shape (Fig. 7A, magenta); a catalytic domain at the short arm of the L shape with the substrate decanal bound at the interface of the catalytic and N-terminal domains (Fig. 7A, ice blue); and, at the long arm of the L shape, a small oligomerization domain (Fig. 7A, green). The closest homologous structures currently available are those reported under PDB accession no. 4KNA (N-succinylglutamate 5-semialdehyde dehydrogenase from Burkholderia thailandensis) and 3JU8, with amino acid sequence identities of 63% and 62%, respectively. The root mean square deviations (RMSDs) for FAldDH with the homologs reported under PDB accession no. 4KNA and 3JU8 are 0.75 and 0.66 Å, respectively, for all C_α atoms (Fig. 7B).

FIG 7.

FAldDH monomer ribbon diagram. (A) Architecture of the three domains in a FAldDH monomer. The cofactor binding domain is in magenta, the catalytic domain is in ice blue, and the oligomerization domain is in green. (B) Superposition of FAldDH onto monomers reported under PDB accession no. 4KNA and 4JU8.

FAldDH dimerization.

The two monomers of FAldDH associate into a tightly intertwined dimer arranged in a head-to-tail fashion (Fig. 8). The extensive interaction buries a total surface area of 3,950 Ų, with a formation energy of −28 kcal/mol calculated by using the program PISA (18). The interface interactions are mainly hydrophobic: 559 nonbonded contacts compared to 10 salt bridges and 54 hydrogen-bonding interactions. The main architecture of the dimer is formed by the NAD⁺ binding domain and catalytic domain, with the extended oligomerization domains docking onto the groove of opposite monomers between the NAD⁺ binding and catalytic domains. The oligomerization domain contributes 3 β-strands to the β-sheet core of the catalytic domain from the other molecule, stabilizing the FAldDH dimer.

FIG 8.

Surface and ribbon representations of the dimerization of FAldDH monomers. The two monomers are shown in green and magenta, respectively. (A) The ethylene glycol bound in the active site and the bound decanal in the substrate binding pocket are shown in blue and dark brown, respectively (PDB accession no. 5U0L). (B) The bound citrate in the active site and NAD⁺ are shown in blue and dark brown sticks, respectively (PDB accession no. 5U0M). Abbreviations: EDO, ethylene glycol; CIT, citrate.

NAD⁺ binding pocket, active site, and substrate binding pocket.

In the FAldDH/NAD⁺ complex structure, each FAldDH monomer is bound with a copy of the NAD⁺ molecule. Similar to other homologous aldehyde dehydrogenase/NAD⁺ complexes, the ADP portion of NAD⁺ has clear electron density, but the ribose sugar connecting the nicotinamide portion has very weak density. However, since the NAD⁺ binding position is known from previous structural studies of related aldehyde dehydrogenases (19) and the nicotinamide has a clear density, intact NAD⁺ was modeled. The extended NAD⁺ cofactor binding pocket is mainly hydrophobic. The ADP moiety of NAD⁺ binds in a pocket from the Rossmann fold N-terminal domain. The 2′-hydroxyl group is in close proximity to C_α of Ser175 (Fig. 9F), which could be relevant to the mechanism of the enzyme distinguishing between NAD⁺ and NADP⁺. The nicotinamide has its ring stacked on the γS and C_β atoms of the catalytic residue Cys282. A citrate binds to the putative active site and interacts with the γS atom of Cys282 through its terminal carboxyl oxygen.

FIG 9.

Ligand binding and interactions in the binding pocket. (A) EDO and the fatty aldehyde substrate tail in the active site/substrate pocket, with the 2mF_o − DF_c electron density map for the decanal tail contoured at 1.0 σ depicted in blue mesh. (B) Ligands observed in the decanal/Maqu_3316 (ligands are decanal and EDO) complex and NAD⁺/Maqu_3316 (ligands are NAD⁺ and CIT) complex structures and in the structure reported under PDB accession no. 2VRO (the ligand is PEG) in the substrate binding pocket (the structures were superimposed to align the substrate binding pockets from each structure). (C) NAD⁺ in the cofactor binding pocket, with the 2mF_o − DF_c electron density map contoured at 1.0 σ depicted in blue mesh. (D) The ligand EDO (blue) and the decanal tail (brown) in the substrate binding pocket and all the amino acids and water molecules (red balls labeled “W”) that are within 4 Å from the ligands. (E) Relative positions of the cofactors, substrates, and substrate or product analogs in the substrate/cofactor binding pockets, with a 90° rotated view of panel B around the vertical axis. (F) Cofactor NAD⁺ interactions with the amino acids forming the cofactor binding pocket.

In the FAldDH/substrate complex, a continuous sausage-like electron density and a separate, smaller, continuous electron density close to γS of Cys282 are observed. Due to the discontinuity of the electron density, a molecule of ethylene glycol (EDO) is modeled near γS of Cys282. The terminal oxygen atom of EDO pointing to γS of Cys282 has a position very similar to that of the oxygen atom of citrate in the FAldDH/NAD⁺ complex. The longer sausage-like electron density was modeled with a nine-atom aliphatic chain, which represents the aliphatic tail of the substrate decanal, which was used for soaking the crystal. Obtaining the complex of the aldehyde dehydrogenase with the linear substrate has been elusive despite a prolonged and intensive effort in the field (20). Compared to the polyethylene glycol (PEG) molecule observed in the structure reported under PDB accession no. 2VRO (Fig. 9B) (19), which was hypothesized to mimic the substrate tail, our results with the aldehyde decanal show that the two molecules have different positions: the PEG molecule in the structure reported under PDB accession no. 2VRO has one end facing the outside of the enzyme, while the decanal in the present structure has both termini in the interior of the enzyme. Having both termini in the interior of the enzyme provides advantages for the enzymatic reaction: the interior of the enzyme could provide a more hydrophobic environment for the substrate and could be more selective for the length of the substrate due to the pocket size limitation.

DISCUSSION

Fatty aldehydes pose a potential problem within the cell, as aldehydes are reactive molecules that can be toxic at elevated concentrations. For organisms that produce these compounds as an intermediate in biosynthetic pathways, managing the levels of internal fatty aldehydes should be very important. For species that fill an important environmental niche and utilize crude oil or other biologically derived lipids as a growth substrate, the oxidation of these aldehydes would be essential for proper function and meeting the energy needs of the cell. Fatty aldehydes have been implicated as a potential intermediate in the biosynthesis of fatty alcohols, although many reports have shown that the enzymes in bacteria that reduce compounds from the fatty acid or activated fatty acid pools (acyl-CoA or acyl-ACP compounds) are also highly efficient at reducing fatty aldehydes and are not likely to release the fatty aldehydes as a reaction intermediate (7, 9, 10, 21), while fatty aldehydes serve as a specific intermediate in the biosynthesis of alkanes in cyanobacteria (16, 17). Interestingly, during a previous analysis of gene transcription in M. aquaeolei VT8 during the accumulation of wax esters, it was found that one specific gene implicated in the potential oxidation of fatty aldehydes (Maqu_3410) was highly transcribed during both exponential growth and wax ester accumulation, even when utilizing citrate as a growth substrate (8). This was somewhat surprising, since FAldDHs are generally assumed to be associated with catabolic processes, and the production of wax esters is believed to be associated with energy storage, involving primarily anabolic processes (12, 13).

Ishige et al. (13) reported the characterization of a FAldDH obtained from Acinetobacter sp. strain M-1 that was able to utilize long-chain fatty aldehydes as a substrate. Since then, there have been few follow-up biochemical studies of these enzymes. Additionally, few reports have detailed the activity of these enzymes with substrates that share similarity with potential native fatty-acid-derived substrates. A BLAST analysis of the genome of M. aquaeolei VT8 revealed quite a few genes that might code for additional FAldDHs. In this work, we cloned and isolated five putative FAldDHs from M. aquaeolei VT8 and also cloned an additional FAldDH from A. baylyi. Of the six proteins that were obtained, only one (Maqu_0607) failed to show any activity with fatty aldehydes. We acknowledge that this lack of activity could be related to the instability of this enzyme during purification or a result of the incorporation of the polyhistidine tag to assist in purification, although the inclusion of the polyhistidine tag did not affect the other five FAldDHs that were isolated and characterized. Several of the other FAldDHs were also prone to large differences in activity for different purifications, although some activity could be rescued from many of these preparations by first degassing the protein preparation under an argon atmosphere and then adding a small amount of β-mercaptoethanol. This approach was not successful for rescuing FAldDH activity with Maqu_0607. Previous reports implicated potential cysteine residues as possible sites that can become oxidized and inhibit enzyme activity (22). We also acknowledge the possibility that this enzyme might be active with a much more specific substrate that was not tested as part of this study.

The genome of M. aquaeolei VT8 contains at least 10 different genes that code for enzymes with the potential to oxidize fatty aldehydes (15). Maqu_3410 (EC 1.2.1—) had the closest similarity score based on a BLAST search using the previously reported fatty aldehyde dehydrogenase from Acinetobacter sp. strain M-1 (13). It was followed closely by another seven genes that had significant similarity scores (E values of <4e−38) and were annotated as various aldehyde dehydrogenases. These included Maqu_0438 (EC 1.2.1.3), Maqu_3572 (also annotated as a coniferyl-aldehyde dehydrogenase [EC 1.2.1.68]), Maqu_3316 (also annotated as a succinylglutamate-semialdehyde dehydrogenase [EC 1.2.1.71]), and Maqu_0607 (EC 1.2.1.3). Each of these genes was cloned and purified as part of this work. Additional genes not characterized included Maqu_3841 (annotated as a betaine aldehyde dehydrogenase [EC 1.2.1.8]), Maqu_3647 (annotated as a succinate semialdehyde dehydrogenase [EC 1.2.1.16]), and Maqu_2133 (annotated as a methylmalonate-semialdehyde dehydrogenase [EC 1.2.1.18]). Several of these genes were also found to be highly transcribed during exponential growth and wax ester accumulation, similar to what was found for Maqu_3410 (8), including Maqu_3316 and Maqu_3572 (21).

Many of the substrates that were selected for this characterization were not easily obtainable from commercial sources. For this reason, we instead chose to utilize a synthetic route to generate these substrates from commercially available fatty alcohols. Based on this work, we were able to characterize the substrate specificity of each of these enzymes with fatty aldehydes that can be derived from the indigenous fatty acid pool that has been characterized for M. aquaeolei VT8 in previous work (3, 8). We acknowledge that the substrates selected in this study do not represent a comprehensive list of potential substrates (or even the most likely natural substrate, based on several of the annotations listed above), but these results provide a measure of activity with small-, medium-, and long-chain fatty aldehydes of various lengths as well as several unsaturated fatty aldehydes. More importantly, all of the assays were performed by the same laboratory under the same conditions so that direct comparisons between these different enzymes can be made with reasonable confidence. Several discrete differences between the five FAldDHs that were isolated in an active form were found (Fig. 2, 4, and 5), and these differences correlate to some degree with the alignments shown in Fig. 3.

The enzyme corresponding to Maqu_3316 yielded diffraction-quality crystals. This enzyme showed a higher degree of specificity for a single substrate (decanal) than did any of the other enzymes characterized in this study with the selection of substrates tested. We note that Maqu_3316 was also annotated as a potential succinylglutamate-semialdehyde dehydrogenase, and succinylglutamate-semialdehyde would have a substrate length similar to that of the substrate decanal, although it would have significant branching groups and additional organic acid functional groups versus the simple straight-chain aldehyde decanal.

The structure solved here with the decanal substrate bound to Maqu_3316 provides evidence of the fatty aldehyde hydrophobic substrate binding pocket for this enzyme. Based on the discontinuity of the electron density, we modeled a portion of the long-chain aldehyde and a molecule of EDO to represent the electron density that was observed. Based on this density, we can define many of the residues that line this potential fatty aldehyde substrate binding pocket in the Maqu_3316 enzyme. Since these residues are expected to mediate substrate selectivity, we performed a more detailed analysis to show the general locations of these residues (Fig. 9D) and show the sequence conservation in the regions surrounding these specific residues from the five enzymes characterized in this study (Fig. 2, bottom) together with the differences in substrate specificity with straight-chain aldehydes (Fig. 2, top). This analysis assumes that each of these enzymes would share significant conservation of the enzyme fold and tertiary structure with the Maqu_3316-derived enzyme. However, strong fold conservation has been the case for this structure and other structures of related aldehyde dehydrogenase enzymes (Fig. 7B), so this should be a reasonable assumption. Based on these results, several specific residues of these enzymes are highlighted as potential targets to alter the potential substrate specificity (Fig. 2). Most profoundly, the His154, Leu155, Ser458, and Ala459 residues seem to be primary determinants of substrate specificity. The Ser458 residue faces the C-8 position of decanal. In Maqu_0438 and Maqu_3572, Ser458 is replaced by a histidine residue, which could interfere with the binding of long-chain aldehydes, while Ala459 is replaced by a glycine residue, which should compensate for the histidine and provide more room and flexibility in the pocket, endowing the enzyme with more tolerance to shorter or longer substrates. In Maqu_3410 and the enzyme reported under RefSeq accession no. WP_004927398, the Ser458 and Ala459 residues are replaced with asparagine and histidine residues, respectively. These substitutions are larger and should create a more rigid binding site, such that shorter substrates would be preferred. Additionally, the His154-Leu155 pair is replaced by leucine and methionine, respectively. This combination may be more flexible, allowing for interactions with longer aliphatic aldehydes. These changes may result in the observed inverted bell-shaped specificity profiles for these two enzymes. Studies to test this hypothesis could be pursued in the future to determine whether mutations introduced into these sites result in changes of substrate specificity.

The two structures that were obtained in this study provide a view for how the FAldDH orients substrates to catalyze the oxidation of a fatty aldehyde to a fatty acid. These structures adopt a typical aldehyde dehydrogenase fold. In each of the two complexes, the substrate or cofactor contains a special “solvent” molecule bound in the active site, which mimics the configuration of the substrate or the product in the catalysis cycle. Figure 10A depicts the four-step enzymatic catalysis. In Fig. 10B, the two structures are superimposed upon one another to bring all of the solvent/ligands into the binding pockets near the active site of the enzyme. The complete decanal molecule is modeled and refined in Fig. 10, and the α-carbon atoms of decanal, EDO, and CIT each overlap in this pocket, which is in close proximity to the catalytic S atom of Cys282. The α-carbon atom of the decanal depicts a longer distance from the S atom. EDO, a substrate analog, occupies the active site, hindering the ability of the aldehyde to bind and blocking decanal access. The ethylene glycol forms two hydrogen bond interactions with the side-chain OH group of Thr438 (3.3 Å) and the S atom of Cys282 (2.2 Å), as shown in Fig. 9D. The C₄ atom of NAD⁺ lies 4.5 Å from the aldehyde oxygen atom of decanal. This distance is sufficient for the C₄ atom of NAD⁺ to accept the H atom from the α-carbon of the decanal during catalysis. Figure 10D shows a citrate molecule occupying the same site. The acid group of the citrate ligand is analogous to the terminal acid of the fatty acid product.

FIG 10.

Enzyme catalytic mechanism and atom configuration of the atoms of the substrate (decanal), the substrate analog EDO/FAldDH complex, and the product analog CIT/FAldDH complex. (A) Two-dimensional diagram of the four-step enzymatic catalysis performed by FAldDH. The substrate and product shown in red represent the two analogs in the present structures. (B) Superposition of the two complex structures showing all of the ligands in the cofactor and substrate binding pockets. The C₄ atom of NAD⁺ that participates in the reduction is labeled and shown in a stick representation. The Cys282 and Glu248 residues that participate in the reaction are also shown. (C) The EDO ligand can be viewed as the aldehyde substrate analog and forms a hydrogen bond with the S atom of Cys282 with a bond distance of 2.2 Å. (D) The CIT ligand with a carboxyl group can be viewed as the acid product analog, which reflects the likely atomic configuration of the product/enzyme complex. The branched atoms are shown in high transparency to highlight the linear portion of the molecule.

Conclusions.

The efforts undertaken in this study combined biochemical and structural studies using a range of different substrates, including several that are not readily available from commercial sources. The results obtained can be used to inform the rational selection of potential residues that might alter the selectivity of these enzymes, which could be used in future biosynthetic approaches to tailor specific reactions. These results also reveal the potential for certain enzymes, such as those annotated as succinylglutamate-semialdehyde dehydrogenases, coniferyl-aldehyde dehydrogenases, or benzaldehyde dehydrogenases, to perform an additional function within the cell, as these enzymes cloned from M. aquaeolei VT8 all exhibited activities similar to one another with a range of small-, medium-, and long-chain aldehydes. Future genetic studies could better identify the roles of these genes and their products in the cell, while mutagenesis studies could probe the features of substrate specificity.

MATERIALS AND METHODS

Strains and materials.

Marinobacter aquaeolei VT8 (ATCC 700491) and Acinetobacter baylyi (ATCC 33305) were obtained from the American Type Culture Collection. Escherichia coli JM109 was obtained from New England BioLabs (Ipswich, MA), while Escherichia coli BL21(DE3) was obtained from Novagen (Madison, WI). All chemicals and reagents were purchased from Sigma-Aldrich (St. Louis, MO) or Fisher Scientific (Pittsburgh, PA) unless otherwise specified. Restriction enzymes and T4 DNA ligase were obtained from New England BioLabs. Coenzymes and fatty aldehydes were sourced from Sigma-Aldrich (St. Louis, MO) or Fisher Scientific (Pittsburgh, PA) or synthesized as follows. All PCRs were performed by using the failsafe PCR enzyme system (Epicentre, Madison, WI).

Synthesis of fatty aldehydes.

Long-chain fatty alcohols were obtained from Nu-Chek Prep (Elysian, MN), and a 100-μl volume of the fatty alcohol was dissolved in methylene chloride (0.1 M) with 10 mol% 2,2,6,6-tetramethylpiperidyl-1-oxyl (TEMPO) and 1:1 equivalents of phenyl iodonium diacetate (PIDA). Aliquots were removed to monitor the progress of the reaction by thin-layer chromatography (TLC) using 4% ethyl acetate (EtOAc) in hexanes (Hex) as the eluent and visualized by using a KMnO₄ stain. The solution was mixed overnight at room temperature in a sealed round-bottom flask. The sample was then analyzed by TLC to determine if it had reached the endpoint, and the volatiles were removed by rotatory evaporation at 35°C under reduced pressure. The sample was then placed under a vacuum in a laboratory HVAC system to remove any remaining volatile material. The mixture of products was then analyzed by using TLC (4% EtOAc in Hex) to obtain R_f values of the products. The product was purified by column chromatography (silica gel, 4% EtOAc in Hex) using vacuum pressure to purify the aldehyde, and the isolated fraction was again dried by rotary evaporation and an HVAC system. The aldehyde was dissolved in CDCl₃ and analyzed by ¹H nuclear magnetic resonance (NMR) spectroscopy to characterize the desired product and assay purity. For tetradecanal (C₁₄H₂₈O), ¹H NMR (400 MHz, CDCl₃) δ 9.76 (t, J = 1.9, 1H), 2.42 (td, J = 7.1, 1.9, 2H), 1.63 (p, J = 7.3, 2H), 1.36 to 1.20 (m, 20H), 0.88 (t, J = 6.8, 3H). For hexadecanal (C₁₆H₃₂O), ¹H NMR (500 MHz, CDCl₃) δ 9.77 (t, J = 1.9, 1H), 2.42 (td, J = 7.4, 1.9, 2H), 1.63 (p, J = 7.4, 2H), 1.37 to 1.19 (m, 24H), 0.88 (t, J = 7.0, 3H). For palmitoleyl aldehyde (C₁₆H₃₀O), ¹H NMR (400 MHz, CDCl₃) δ 9.76 (t, J = 1.9, 1H), 5.40 to 5.29 (m, 2H), 2.42 (td, J = 7.4, 1.9, 2H), 2.01 (q, J = 6.2, 4H), 1.63 (p, 7.2, 2H), 1.40 to 1.20 (m, 16H), 0.88 (t, J = 6.7, 3H). For oleyl aldehyde (C₁₈H₃₄O), ¹H NMR (400 MHz, CDCl₃) δ 9.76 (t, J = 1.9, 1H), 5.40 to 5.29 (m, 2H), 2.42 (td, J = 7.4, 1.9, 2H), 2.01 (q, J = 6.0, 4H), 1.63 (p, J = 7.2, 2H), 1.38 to 1.20 (m, 20H), 0.88 (t, J = 6.7, 3H). Once these aldehydes were obtained, they were stored frozen under an argon atmosphere until required for testing.

Plasmid constructions.

The gene for a putative fatty aldehyde dehydrogenase (Maqu_3410) from M. aquaeolei VT8 was cloned by PCR with primers BBP801 (5′-GACTACCATGGAATCTATGCACAACCCGGACAAGAAGGCTCCGTTG-3′) and BBP802 (5′-GTCATTCTAGAATCAGAAGAACCCGAGAGGGTTGGTGTCGTAGCTG-3′) and then digested with the restriction enzymes NcoI and XbaI (underlined in primer sequences for clarity) and ligated into plasmid pBB114, a pUC19 derivative containing kanamycin in place of ampicillin resistance (8). Following sequencing to confirm the gene, this DNA segment was shuttled, using the same restriction enzymes, into a pET-19b derivative plasmid to create plasmid pETMFA, which includes an 8× polyhistidine tag at the N terminus.

The gene for a second putative fatty aldehyde dehydrogenase (Maqu_0438) from M. aquaeolei VT8 was inadvertently cloned by PCR with primers BBP1778 (5′-NNNGGATCCATGGACAGATTGCTAGTCTGGCGGAGGTTGG-3′) and BBP1779 (5′-NNNAAGCTTCAGCCCAGAATACGAACAGCAGCATCTATG-3′) and then digested with the restriction enzymes BamHI and HindIII and ligated into plasmid pBB114. Two mistakes found in the primer regions of this construct were repaired by site-specific mutagenesis (Stratagene method; Agilent Technologies, Santa Clara, CA). Following sequencing to confirm the gene, this DNA segment was shuttled, using NcoI and HindIII, into a pET-19b derivative plasmid to create plasmid pPCRMFA9, which includes an 8× polyhistidine tag at the N terminus.

The gene for a third putative fatty aldehyde dehydrogenase (Maqu_0607) from M. aquaeolei VT8 was also cloned by PCR with primers BBP1778 and BBP1779 and then digested with the restriction enzymes BamHI and HindIII and ligated into plasmid pBB114. Site-specific mutagenesis was used to insert a silent mutation that removed a second NcoI site within the gene. Following sequencing to confirm the gene, this DNA segment was shuttled, using NcoI and HindIII, into a pET-19b derivative plasmid to create plasmid pPCRMFA10, which includes an 8× polyhistidine tag at the N terminus.

The gene for a fourth putative fatty aldehyde dehydrogenase (Maqu_3572) from M. aquaeolei VT8 was cloned by PCR with primers BBP2319 (5′-NNNAAGCTTCAGCGAATAAACAGCTTATACACCAACC-3′) and BBP2320 (5′-NNNATGCATCACCACCATCATCACGGTGCCACCGTCGTCCAGCTCACC-3′) and then digested with the restriction enzymes NsiI and HindIII and ligated into plasmid pBB114. Following sequencing to confirm the gene, this DNA segment was shuttled, using NdeI and HindIII, into a pET-19b derivative plasmid to create plasmid pPCRMFA36, which includes a 6× polyhistidine tag at the N terminus.

The gene for a fifth putative fatty aldehyde dehydrogenase (Maqu_3316) from M. aquaeolei VT8 was cloned by PCR with primers BBP2343 (5′-NNNAAGCTTGGTTGGCCACCGCATTGACGTTGG-3′) and BBP2344 (5′NNNTCTAGAGTAACCACCATACCCATGAACTGCATCG-3′) and then digested with the restriction enzymes HindIII and XbaI and ligated into plasmid pBB114. A polyhistidine tag was then added to this gene by using primers BBP2420 (5′-NNNGGTACCATATGCATCACCACCATCACCATGCAAACCTGACAGGCAATGTGTACATC-3′) and BBP2421 (5′-GATCCCCGGGTACCGAG CTCGAATTCACTG-3′) and then digesting the product with KpnI and ligation. Following sequencing to confirm the gene, this DNA segment was shuttled, using NdeI and HindIII, into a pET-19b derivative plasmid to create plasmid pPCRMFA47, which includes a 6× polyhistidine tag at the N terminus.

The gene for a sixth putative fatty aldehyde dehydrogenase (RefSeq accession no. WP_004927398) from Acinetobacter baylyi was cloned by PCR with primers BBP2389 (5′-NNNATGCATCACCACCATCATCACCGTTATATCGATCCTAATCAACCTGGCTC-3′) and BBP2390 (5′-NNNAAGCTTAGAAGAAGCCCATTGGTTTTGTTGAATAAC-3′) and then digested with the restriction enzymes NsiI and HindIII and ligated into plasmid pBB114. Following sequencing to confirm the gene, this DNA segment was shuttled, using NdeI and HindIII, into a pET-19b derivative plasmid to create plasmid pPCRMFA48, which includes a 6× polyhistidine tag at the N terminus.

Protein purification.

Completed plasmids were transformed into E. coli BL21(DE3) for the expression of the protein. Cells were grown in Miller's lysogeny broth (LB) at 30°C and induced with isopropyl-β-d-thiogalactopyranoside (IPTG) (50 mg liter⁻¹) for 2 h, when they reached an optical density at 600 nm of 0.6. Cells were harvested by centrifugation at 7,000 × g for 7 min. Protein was extracted in a manner similar to what was described previously, using metal affinity chromatography followed by a desalting column for equilibration in a final buffer containing 50 mM potassium phosphate (pH 7.0) and 300 mM NaCl (23). Final isolated fractions were pooled and tested for purity by running the fractions on SDS-PAGE gels. The protein was either flash frozen in liquid nitrogen or stored at 4°C.

Fatty aldehyde dehydrogenase activity assays.

Enzyme activity assays were conducted with quartz cuvettes with a 1-cm path length using a Varian Bio 50 UV-visible (UV-Vis) spectrophotometer equipped with temperature control. Each assay mixture contained 900 μl of reaction buffer (25 mM glycine and 50 mM potassium phosphate [pH 9.6]). Aldehydes were diluted in isopropanol such that 200 nmol was added to the cuvette. NAD⁺ was then added from a concentrated stock solution to bring the final concentration to 1.5 μmol. The various components were mixed thoroughly with a pipette before being placed into the spectrophotometer. Samples were monitored by measuring the absorbance at 340 nm for 1 min to confirm a stable baseline before the addition of the enzyme (10 to 100 μg) to initiate the reaction. The production of NADH was determined by correlating the absorbance change to the extinction coefficient for NADH (6,220 M⁻¹ cm⁻¹ at 340 nm). Initial rates of reaction were calculated by exporting the raw data to Microsoft Excel (Microsoft, Redmond, WA) to convert the absorbance per minute to specific activity rates. NADP⁺-based assays were performed according to the same procedure but with NADP⁺ substituted for NAD⁺. All calculations are based on data from a minimum of 3 replicates (n ≥ 3). Saturated aliphatic aldehyde substrates tested included acetaldehyde (C₂), butanal (C₄), hexanal (C₆), octanal (C₈), decanal (C₁₀), dodecanal (C₁₂), tetradecanal (C₁₄), and hexadecanal (C₁₆). Unsaturated aliphatic aldehyde substrates tested included palmitoleyl aldehyde (C_16:1) and oleyl aldehyde (C_18:1). Quantification of enzyme concentrations used for calculating specific activities were based on the absorbance of the protein at 280 nm and the extinction coefficient calculated with the ExPaSy Protein Parameters algorithm (see http://web.expasy.org/protparam/) (24).

Sequence comparisons.

Enzymes similar to the original fatty aldehyde dehydrogenase from Acinetobacter sp. strain M-1 (13) and the genes cloned here were used to perform a BLAST search. Enzymes that shared high identity and had three-dimensional (3D) models in the Protein Data Bank were compared to our selection of enzymes. The sequences of the enzyme reported under RefSeq accession no. WP_004927398, Maqu_3316, Maqu_0438, Maqu_3572, Maqu_3410, and six closely related sequences were compared by using Multalin (http://multalin.toulouse.inra.fr/multalin/) (25). Sequence alignments were generated along with a hierarchical tree of similarity.

Crystallization and structure determination of a FAldDH.

Purified Maqu_3316 at 36 mg/ml was screened for crystallization by the sitting-drop vapor diffusion method at the Nanoliter Crystallization Facility at the Department of Biochemistry, Molecular Biology and Biophysics, University of Minnesota, using the CrystalTrak system (Rigaku, Tokyo, Japan). In total, 10 screens, each consisting of 96 unique conditions, were set up. Crystals appeared after 3 weeks under multiple conditions. Crystals were transferred into cryoprotectant solutions containing the corresponding reservoir solution supplemented with 20% (vol/vol) ethylene glycol. For ligand soaking, the cryoprotectant solution contained 20 mM decanal or NAD⁺. The typical soaking time was 1 to 2 h. The crystals were then flash frozen in liquid nitrogen.

Crystals were screened at the Advanced Photon Source Northeastern Collaborative Access Team beamlines (24-ID-C and 24-ID-E). The best-diffracting crystals were the ones from the MCSG_2 screen (Anatrace, Maumee, OH) (well ID F8, 0.2 M NH₄H₂PO₄, 50% [vol/vol] 2-methyl-2,4-pentanediol, 0.1 M Tris [pH 8.5]) and the PEGRx HT screen (Hampton Research, Aliso Viejo, CA) (well ID F10 containing 0.2 M ammonium citrate, 20% [wt/vol] PEG 2,000 monomethyl ether, 0.1 M imidazole [pH 7.0]). The collected data were processed with XDS (26). The Matthews coefficient calculation indicated that there would likely be a dimer of Maqu_3316 in an asymmetric unit. Using the monomer of succinylglutamic semialdehyde dehydrogenase from Pseudomonas aeruginosa (PDB accession no. 3JU8) as a search model, molecular replacement by PHASER (27) located two copies of the FAldDH monomers in the asymmetric unit. Subsequent iterative refinement with PHENIX suite (28) and model inspection and building using COOT (29) resulted in final R_work/R_free values of 22.62% and 26.72% for the FAldDH/substrate complex and 21.78% and 27.30% for the FAldDH/NAD⁺ complex, respectively. A summary of the data collection and refinement statistics is shown in Table 2. Ramachandran analysis shows that 96.1%, 3.9%, and 0% of the protein residues are in the most favored, allowed, and disallowed regions for the FAldDH/substrate complex and that 95.9%, 4.1%, and 0% of the protein residues are in the most favored, allowed, and disallowed regions for the FAldDH/NAD⁺ complex, respectively. Molecular graphic images were produced by using PYMOL (http://www.pymol.org/).

TABLE 2.

Data collection and refinement statistics^a

Parameter	Value(s) for complex:
	FAldDH/substrate	FAldDH/NAD+
Data collection statistics
Resolution range (Å)	92.04–2.29 (2.37–2.29)	78.33–3.08 (3.29–3.08)
Space group	P41212	P41212
Unit cellb (Å)	98.71, 98.71, 254.73	99.38, 99.38, 254.60
Total no. of reflections	841,693 (79,325)	178,295 (31,311)
No. of unique reflections	57,590 (4,945)	24,631 (4,349)
Multiplicity	14.6 (14.3)	7.2 (7.2)
Completeness (%)	97.41 (86.95)	99.90 (99.90)
I/σI	14.82 (0.84)	6.30 (0.70)
Rmerge	0.176 (4.78)	0.371 (2.64)
Rmeas	0.182 (4.96)	0.399 (2.84)
Rp.i.m.	0.04713 (1.282)	0.1446 (1.035)
CC1/2	0.999 (0.563)	0.991 (0.505)
Refinement statistics
No. of reflections	56,699 (4,937)	23,468 (2,126)
No. of reflections for Rfree	2,742 (235)	1,139 (98)
Rwork (%)	22.62 (42.82)	21.78 (40.43)
Rfree (%)	26.72 (46.56)	27.30 (43.39)
No. of nonhydrogen atoms	7,402	7,475
Macromolecules	7,350	7,350
Ligands	4	92
Solvent	48	33
No. of protein residues	975	975
RMSD
Bond length (Å)	0.002	0.005
Bond angle (°)	0.61	0.72
Ramachandran plot (%)
Favored regions	96.1	95.9
Allowed regions	3.9	4.1
Outliers	0	0
Avg B factor (Å2)	73.92	86.09
Macromolecules	73.98	85.84
Ligands	88.34	121.80
Solvent	62.38	43.41

^aStatistics for the highest-resolution shell are shown in parentheses.

^bValues shown are for dimensions a, b, and c.

Accession number(s).

The structure factors and coordinates have been deposited in the Protein Data Bank (30) under accession no. 5U0L and 5U0M, respectively.