Human Aldehyde Dehydrogenase Genes: Alternatively-Spliced Transcriptional Variants and Their Suggested Nomenclature

Abstract

OBJECTIVE

The human aldehyde dehydrogenase (ALDH) gene superfamily consists of 19 genes encoding enzymes critical for NAD(P)⁺-dependent oxidation of endogenous and exogenous aldehydes, including drugs and environmental toxicants. Mutations in ALDH genes are the molecular basis of several disease states (e.g. Sjögren-Larsson syndrome, pyridoxine-dependent seizures, and type II hyperprolinemia) and may contribute to the etiology of complex diseases such as cancer and Alzheimer’s disease. The aim of this nomenclature update was to identify splice transcriptional variants principally for the human ALDH genes.

METHODS

Data-mining methods were used to retrieve all human ALDH sequences. Alternatively-spliced transcriptional variants were determined based upon: a) criteria for sequence integrity and genomic alignment; b) evidence of multiple independent cDNA sequences corresponding to a variant sequence; and c) if available, empirical evidence of variants from the literature.

RESULTS AND CONCLUSION

Alternatively-spliced transcriptional variants and their encoded proteins exist for most of the human ALDH genes; however, their function and significance remain to be established. When compared with the human genome, rat and mouse include an additional gene, Aldh1a7, in the ALDH1A subfamily. In order to avoid confusion when identifying splice variants in various genomes, nomenclature guidelines for the naming of such alternative transcriptional variants and proteins are recommended herein. In addition, a web database (www.aldh.org) has been developed to provide up-to-date information and nomenclature guidelines for the ALDH superfamily.

Introduction

Aldehydes are highly reactive compounds capable of exerting a variety of toxic cellular events including adduct formation with DNA and proteins. Endogenous aldehydes are formed during the metabolism of numerous compounds including alcohols, amino acids, biogenic amines, vitamins, steroids and lipids. Exogenous aldehydes are often generated from the biotransformation of drugs and environmental agents [1, 2]. The mammalian ALDH gene superfamily encodes a group of evolutionarily-related sequences whose protein products all have pyridine nucleotide-dependent oxidation activity catalyzing the irreversible oxidation of aldehydic substrates to their corresponding carboxylic acids [3-5].

Although many ALDH enzymes display broad substrate specificity and oxidize a variety of aliphatic and aromatic aldehydes, others retain unique substrate preferences. In addition to their primary role in aldehyde oxidation, many ALDH enzymes possess multiple catalytic and non-catalytic functions. For example, ALDH1A1, ALDH2, ALDH3A1 and ALDH4A1 catalyze ester hydrolysis; in the case of ALDH2, this hydrolytic activity has been implicated in the bioactivation of nitroglycerin to nitric oxide [6, 7]. ALDH1A1 is capable of binding androgens, cholesterol, thyroid hormone and flavopyridol whereas ALDH2 has been identified as an acetaminophen-binding protein [4, 8]. ALDH proteins have been hypothesized to play a critical role in cellular homeostasis by maintaining redox balance [9]. For example, ALDH enzymes contribute to the antioxidant capacity of a cell by generating NAD(P)H, which can be used for the regeneration of reduced glutathione (GSH). Furthermore, it has been proposed that ALDH3A1 may scavenge hydroxyl radicals via reduction of its cysteine and methionine thiol groups [10, 11]. The ALDH proteins not only differ with regard to their catalytic/non-catalytic properties and tissue distribution but also in relation to their sensitivity to inhibitors, suppressors and inducers.

The clinical importance of ALDH enzymes is evident from the observation that mutations and polymorphisms in ALDH genes (leading to loss of function) are associated with distinct phenotypes in humans [8, 12]—including Sjögren-Larsson syndrome [13], type II hyperprolinemia [14], γ-hydroxybutyric aciduria [15], pyridoxine-dependent seizures [16], hyperammonemia [17], alcohol-related diseases [18], cancer [19] and late-onset Alzheimer’s disease [20]. Aside from the clinical phenotypes associated with mutations in ALDH genes, knockout mouse models have suggested a crucial role of ALDH enzymes in physiological functions and processes, such as embryogenesis and development [21, 22] as well as protection against oxidative stress [23].

A growing body of evidence supports the expression of alternatively-spliced transcriptional variants for many of the ALDH genes. However, the spatiotemporal factors affecting this expression (as well as their physiologic roles) remain unclear. In the present paper, we describe and classify alternatively-spliced transcript products within the human ALDH gene superfamily. These alternatively-spliced variants were identified within the molecular sequence libraries from the National Center for Biotechnology Information (NCBI) and the European Bioinformatics Institute (EBI) and classified in accordance with recommended nomenclature guidelines for the naming of such alternative transcriptional variants and their proteins.

To assist readers and to provide a detailed resource for the ALDH gene superfamily, an ALDH database is located on the web at www.aldh.org. Extensive information for each ALDH gene found in human, other animals, archaebacteria, eubacteria, fungi, plant, and yeast genomes is available—including information on the current practices of the ALDH nomenclature system. There are also links to other informational databases and programs for analyzing protein and DNA sequences, such as those maintained by NCBI. Furthermore, graphical and tabular representation of all transcriptional variants and corresponding proteins described in this present report are available at www.aldh.org for visual reference.

Methods

Data mining was employed to identify new (and existing), putatively-functional ALDH protein-coding sequences and relevant information for the genes, transcripts, and corresponding proteins of mammalian genomes from the human, mouse, rat, rhesus monkey, chimpanzee, cow, dog, rabbit and opossum. Transcript and peptide sequence orthologs were identified utilizing the Basic Local Alignment Search Tool (BLAST) program [24]. Multiple sequence alignments using Clustal W [25] and T-Coffee [26] were used to compare and catalog ALDH genes across species. We also created an evolutionary dendrogram of known human, mouse and rat ALDH sequences (Figure 1).

Figure 1.

Dendrogram illustrating the evolutionary relationship of ALDH protein sequences from human, mouse, and rat. Accession numbers for ALDH sequences are provided at www.aldh.org.

Sequences for all transcript and peptide translations of accession identification numbers referenced within are available from the NCBI and European Molecular Biology Laboratory (EMBL)-EBI databases. These entities were analyzed for sequence integrity and genomic alignment based upon the most recent build assemblies available from these institutes at the time of this writing. Transcript sequences were aligned with their corresponding genomic assembly using our proprietary SAST alignment software (2009, W. Black and V. Vasiliou, manuscript in preparation) and confirmed with NCBI’s Splign utility [27].

The structural integrity of all transcript sequences was determined to have a coding sequence beginning with a 5’methionine initiation codon (ATG) and a 3’ termination codon (TGA, TAG or TAA). Translation of this coding sequence was then analyzed to confirm that the corresponding reading frame retained an ALDH peptide domain according to the Hidden Markov Model (HMM) for this domain, termed “aldedh”, available from Pfam [28]. Alternatively-spliced transcriptional variants (described herein) were determined based upon: a) criteria for sequence integrity and genomic alignment; b) evidence of multiple independent cDNA sequences corresponding to a variant sequence; and c) if available, empirical evidence of variants from the literature. Multiple independent cDNA sequences that were associated with a particular variant were considered indicative of a potential alternatively-spliced transcriptional variant; unique sequences were not described but were shelved for further analysis and data support.

The identification of splice transcripts and the resulting proteins raises the issue of nomenclature for these entities within existing and future literature, as they are identified in various genomes. In keeping with the Human Gene Nomenclature Guidelines, alternatively-spliced transcriptional variants and corresponding proteins are denoted by a “_v” symbol followed by a number indicating the variant (e.g. ALDH3A1_v2). Manuscripts describing an ALDH entity subject to alternative splicing should clearly state the variant being studied. In this regard, different alternative transcriptional variants and corresponding proteins may prove to have vastly different properties and functionalities. In the human genome, evidence for alternative transcripts exists for most of the 19 ALDH genes—with the exception of ALDH1B1, ALDH2, ALDH7A1 and ALDH9A1.

The ALDH-like Clan and the Mammalian ALDH Gene Superfamily

The ALDH gene superfamily is included in the ALDH-like clan (Pfam CL0099) which consists of four members; the ALDH gene superfamily (Pfam “Aldedh”), a family of uncharacterized proteins from Drosophila melanogaster (Pfam DUF1487; PF07368), a histidinol dehydrogenase family (Pfam “Histidinol_dh”; PF00815), and an acyl-CoA reductase family (Pfam “LuxC”; PF05893). Members of the ALDH gene superfamily are widely expressed among eukaryotes and prokaryotes. Analysis of mammalian genomes has revealed the presence of 19 or 20 ALDH gene orthologs per species. A clustering dendrogram of the human, mouse and rat ALDHs is shown in Figure 1. To date, 19 putatively functional ALDH genes exist in the human genome and a brief description of the function of these gene products is provided in Table 1.

Table 1.

Human ALDH genes and gene products

Gene	Protein Description
ALDH1A1	ALDH1A1 is a cytosolic enzyme that oxidizes retinal, acetaldehydes and 3-deoxyglucosone (a product of protein deglycation and a potent glycating agent).
ALDH1A2	ALDH1A2 is a cytosolic enzyme that is integrally involved in the oxidation of retinal to retinoic acid during embryonic development. Aldh1a2(-/-) mice are embryolethal.
ALDH1A3	ALDH1A3 is a cytosolic retinaldehyde-metabolizing enzyme.
ALDH1B1	ALDH1B1 is a mitochondrial enzyme that metabolizes acetaldehyde.
ALDH1L1	ALDH1L1 is a fusion protein comprising three domains: a formyl transferase domain at the amino terminal, a centrally-located formyltransferase carboxyl terminal domain and an aldehyde dehydrogenase domain at its carboxyl terminal (Figure 2).
ALDH1L2	ALDH1L2 shares ≈73% identity with ALDH1L1; no functional data have been reported for this protein.
ALDH2	ALDH2 is a mitochondrial enzyme involved in the oxidation of acetaldehyde and the metabolites of dopamine and norepinephrine, DOPAL and DOPEGAL, respectively.
ALDH3A1	ALDH3A1 is a multifunctional enzyme that plays a significant role in the cellular response to oxidative stress.
ALDH3A2	ALDH3A2 is a microsomal enzyme that oxidizes medium to long-chain fatty aldehydes.
ALDH3B1	ALDH3B1 is a cytosolic protein that oxidizes medium- and long-chain saturated and unsaturated aliphatic aldehydes.
ALDH3B2	ALDH3B2 is a putative ALDH with no functional data available.
ALDH4A1	ALDH4A1 catalyzes the irreversible conversion of Δ1-pyrroline-5-carboxylate (derived from either proline or ornithine) to glutamate, necessary to connect the urea cycle with the tricarboxylic acid cycle.
ALDH5A1	ALDH5A1 is the succinate semialdehyde dehydrogenase involved in the last step of GABA catabolism, converting GABA to succinate semialdehyde.
ALDH6A1	ALDH6A1 is the methylmalonate semialdehyde dehydrogenase that catalyzes the irreversible oxidative decarboxylation of malonate and methylmalonate semialdehydes to acetyl- and propionyl-CoA, respectively.
ALDH7A1	ALDH7A1 metabolizes α-aminoadipic semialdehyde, generated during lysine catabolism.
ALDH8A1	ALDH8A1 appears to be involved in 9-cis-retinoic acid biosynthesis.
ALDH9A1	ALDH9A1 catalyzes the oxidation of γ-aminobutyraldehyde and betaine aldehyde, a γ-trimethylaminobutyraldehyde.
ALDH16A1	No functional information exists in the literature for this enzyme.
ALDH18A1	ALDH18A1 is a bi-functional ATP- and NAD(P)H-dependent mitochondrial inner-membrane protein having both γ-glutamyl kinase and γ-glutamyl phosphate reductase activities

ALDH1 Family

The ALDH1 family consists of six human ALDH genes: ALDH1A1, ALDH1A2, ALDH1A3, ALDH1B1, ALDH1L1 and ALDH1L2. The genomes of Rattus norvegicus (rat) and Mus musculus (mouse) contain an additional gene, Aldh1a7 that is 92% identical to mouse Aldh1a1. Therefore, the rodent Aldh1a7 very likely arose as a gene duplication event after the mammalian radiation ~70 million years ago (MYA) and then became fixed in the genome before the rat-mouse divergence ~17 MYA.

ALDH1A1

Two transcriptional variants identified for the human ALDH1A1 gene consist of 13 and 8 exons for the consensus ALDH1A1_v1 and ALDH1A1_v2, respectively (Table 2). Relative to the native ALDH1A1_v1, the ALDH1A1_v2 variant lacks the 3’ end of exon 7, a portion of the 5’ and 3’ ends of exon 9, and is missing exons 8, 10, 11, 12 and 13. This translates to a protein splice-variant missing 271 amino acids from the COOH-terminus, relative to the native form. Pfam analysis revealed this protein splice-variant retains an ALDH peptide domain—although truncated. The predicted active-site cysteine and glutamate residues of the primary variant ALDH1A1_v1 at positions 303 and 269, respectively, are not apparent within the ALDH1A1_v2 variant, strongly suggesting that this protein likely has no ALDH activity.

Table 2.

Human ALDH Alternative Transcripts

Gene	Transcript	Exons	Clones*	Transcript Accession ‡	Peptide	Peptide Accession ‡	Length (amino acids)	M.W. (kDa)
ALDH1A1
	ALDH1A1	13	236	NM_000689	ALDH1A1	NP_000680	501	54.7
	ALDH1A1_v2	8	16	ENST00000376939	ALDH1A1_v2	ENSP00000366138	230	25.3
ALDH1A2
	ALDH1A2	13	135	NM_003888	ALDH1A2	NP_003879	518	56.5
	ALDH1A2_v2	12	5	NM_170696	ALDH1A2_v2	NP_733797	480	52.9
	ALDH1A2_v3	11	5	NM_170697	ALDH1A2_v3	NP_733798	422	46.0
	ALDH1A2_v4	12	5	ALDH1A2.cApr07	ALDH1A2_v4	ALDH1A2.cApr07	384	42.4
ALDH1A3
	ALDH1A3	13	153	NM_000693	ALDH1A3	NP_000684	512	55.9
	ALDH1A3_v2	10	158	ENST00000346623	ALDH1A3_v2	ENSP00000343294	416	45.4
ALDH1B1
	ALDH1B1	2	213	NM_000692	ALDH1B1	NP_000683	517	57.2
ALDH1L1
	ALDH1L1	23	190	NM_012190	ALDH1L1	NP_036322	902	98.6
	ALDH1L1_v2	23	1	ENST00000273450	ALDH1L1_v2	ENSP00000273450	912	99.7
	ALDH1L1_v3	22	N.A.	ENST00000393431	ALDH1L1_v3	ENSP00000377081	505	55.3
	ALDH1L1_v4	7	7	ALDH1L1.hApr07	ALDH1L1_v4	ALDH1L1.hApr07	333	36.4
	ALDH1L1_v5	6	6	ALDH1L1.jApr07	ALDH1L1_v5	ALDH1L1.jApr07	259	28.5
ALDH1L2
	ALDH1L2	23	10	NM_001034173	ALDH1L2	NP_001029345	923	101.6
	ALDH1L2_v2	11	37	ALDH1L2.cApr07	ALDH1L2_v2	ALDH1L2.cApr07	378	41.4
	ALDH1L2_v3	22	34	ALDH1L2.aApr07	ALDH1L2_v3	ALDH1L2.aApr07	810	89.1
ALDH2
	ALDH2	13	222	NM_000690	ALDH2	NP_000681	517	56.3
ALDH3A1
	ALDH3A1	11	325	NM_000691	ALDH3A1	NP_000682	453	50.4
	ALDH3A1_v2	9	63	ALDH3A1.aApr07	ALDH3A1_v2	ALDH3A1.aApr07	570	61.6
	ALDH3A1_v3	11	44	ALDH3A1.dApr07	ALDH3A1_v3	ALDH3A1.dApr07	452	50.3
	ALDH3A1_v4	9	31	ALDH3A1.hApr07	ALDH3A1_v4	ALDH3A1.hApr07	323	35.7
	ALDH3A1_v5	8	N.A.	ENST00000333946	ALDH3A1_v5	ENSP00000334590	570	61.5
	ALDH3A1_v6	10	1	ENST00000395555	ALDH3A1_v6	ENSP00000378923	389	43.3
	ALDH3A1_v7	10	N.A.	ALDH3A1.eApr07	ALDH3A1_v7	ALDH3A1.eApr07	380	41.9
ALDH3A2
	ALDH3A2	10	191	NM_000382	ALDH3A2	NP_000373	485	54.9
	ALDH3A2_v2	11	18	NM_001031806	ALDH3A2_v2	NP_001026976	508	57.5
	ALDH3A2_v3	11	11	ENST00000395575	ALDH3A2_v3	ENSP00000378942	485	54.8
	ALDH3A2_v4	10	N.A.	ENST00000404114	ALDH3A2_v4	ENSP00000385699	508	57.6
	ALDH3A2_v5	7	38	ALDH3A2.eApr07	ALDH3A2_v5	ALDH3A2.eApr07	292	33.0
	ALDH3A2_v6	3	5	ALDH3A2.lApr07	ALDH3A2_v6	ALDH3A2.lApr07	97	10.9
ALDH3B1
	ALDH3B1	10	45	NM_000694	ALDH3B1	NP_000685	468	51.7
	ALDH3B1_v2	9	18	NM_001030010	ALDH3B1_v2	NP_001025181	431	47.5
	ALDH3B1_v3	9	98	ALDH3B1.dApr07	ALDH3B1_v3	ALDH3B1.dApr07	248	27.6
	ALDH3B1_v4	7	3	ALDH3B1.eApr07	ALDH3B1_v4	ALDH3B1.eApr07	223	24.7
	ALDH3B1_v5	9	4	ALDH3B1.kApr07	ALDH3B1_v5	ALDH3B1.kApr07	88	9.6
ALDH3B2
	ALDH3B2	10	89	NM_000695	ALDH3B2	NP_000686	385	42.4
	ALDH3B2_v2	10	101	NM_001031615	ALDH3B2_v2	NP_001026786	385	42.4
	ALDH3B2_v3	9	2	ALDH3B2.cApr07	ALDH3B2_v3	ALDH3B2.cApr07	357	39.3
ALDH4A1
	ALDH4A1	15	203	NM_003748	ALDH4A1	NP_003739	563	61.7
	ALDH4A1_v2	16	2	NM_170726	ALDH4A1_v2	NP_733844	563	61.7
	ALDH4A1_v3	14	N.A.	ENST00000375335	ALDH4A1_v4	ENSP00000364484	547	59.8
	ALDH4A1_v4	8	N.A.	ENST00000375334	ALDH4A1_v3	ENSP00000364483	195	21.2
	ALDH4A1_v5	9	2	ALDH4A1.eApr07	ALDH4A1_v5	ALDH4A1.eApr07	195	21.2
ALDH5A1
	ALDH5A1	10	216	NM_001080	ALDH5A1	NP_001071	535	57.2
	ALDH5A1_v2	11	10	NM_170740	ALDH5A1_v2	NP_733936	548	58.6
	ALDH5A1_v3	4	5	ALDH5A1.cApr07	ALDH5A1_v3	ALDH5A1.cApr07	172	18.5
ALDH6A1
	ALDH6A1	12	427	NM_005589	ALDH6A1	NP_005580	535	57.8
	ALDH6A1_v2	7	8	ALDH6A1.bApr07	ALDH6A1_v2	ALDH6A1.bApr07	293	31.6
	ALDH6A1_v3	5	3	ALDH6A1.cApr07	ALDH6A1_v3	ALDH6A1.cApr07	179	19.6
	ALDH6A1_v4	4	5	ALDH6A1.jApr07	ALDH6A1_v4	ALDH6A1.jApr07	117	12.7
ALDH7A1
	ALDH7A1	18	187	NM_001182	ALDH7A1	NP_001173	511	55.2
ALDH8A1
	ALDH8A1	7	68	NM_022568	ALDH8A1	NP_072090	487	53.2
	ALDH8A1_v2	6	3	NM_170771	ALDH8A1_v2	NP_739577	433	47.1
ALDH9A1
	ALDH9A1	11	246	NM_000696	ALDH9A1	NP_000687	518	56.1
ALDH16A1
	ALDH16A1	17	153	NM_153329	ALDH16A1	NP_699160	802	84.9
	ALDH16A1_v2	15	1	ALDH16A1andFLT3LG.cApr07	ALDH16A1_v2	ALDH16A1andFLT3LG.cApr07	292	31.6
ALDH18A1
	ALDH18A1	18	434	NM_002860	ALDH18A1	NP_002851	795	87.1
	ALDH18A1_v2	18	11	NM_001017423	ALDH18A1_v2	NP_001017423	793	86.9

^*Number of clones, as provided by the NCBI-AceView database.

^‡Accession identification numbers from NCBI – GenBank have the format “NM_…”, “NP_…”, “XM_…”, or “XP_…”; from EBI – Ensembl have the format “ENS…”; and from NCBI – AceView have the format “ALDH#X#.xApr07”.

ALDH1A2

Four distinct human ALDH1A2 transcriptional variants have been identified (Table 2). The consensus ALDH1A2 variant, ALDH1A2_v1, represents the longest and most prevalent transcript and protein. Interestingly, intron 1 of both ALDH1A2_v1 and ALDH1A2_v2 is quite large (51.4 kb). ALDH1A2_v2 lacks the exon 7 segment present in the primary variant ALDH1A2_v1. Exon 7 is within the coding region of the transcript; the lack of this segment translates to a shorter protein. Variant ALDH1A2_v3, a derivative of ALDH1A2_v1, lacks exons 1 and 2 of ALDH1A2_v1. Relative to ALDH1A2_v1, the first exon of ALDH1A2_v3 contains a distinct 5’-untranslated region (UTR) comprising an additional 15-bp segment upstream of exon 3. The resulting protein variant has a shorter NH₂-terminus in comparison to the major variant ALDH1A2_v1. A fourth variant identified within the sequence databases, ALDH1A2_v4, is a derivative of the ALDH1A2_v2 variant and lacks the 114-bp exon 7 of ALDH1A2_v1. This variant, however, utilizes an alternate exon 1 leading to a modified 5’ coding region.

ALDH1A3

The human ALDH1A3 gene includes two variant transcripts (Table 2). Although only a single transcript is reported by RefSeq in the NCBI Entrez Gene database (GeneID 220), a second variant, ALDH1A3_v2 is readily apparent according to cDNA evidence (Table 2) and as described by EMBL-EBI’s Ensembl (ENST00000346623). The ALDH1A3_v2 variant transcript lacks exons 4, 5, and 6—compared with ALDH1A3_v1—and encodes a splice-variant that is missing an internal segment within the ALDH peptide domain 5’ to the predicted cysteine and glutamate residues in the active-site.

Aldh1a7

Mouse Aldh1a7 most closely resembles an ancestral Aldh1a1 homolog when examined using evolutionary divergence (Figure 1). Comparing Aldh1a7 exon segments to other mammalian genomes using BLAST analysis does not produce significant correlations, suggesting speciation is limited. Details of alternatively-spliced transcriptional variants for the mouse and rat are beyond the scope of this manuscript. However, preliminary evidence suggests there are two transcriptional variants within NCBI’s AceView database accession identification numbers Aldh1a7.aSep07 and Aldh1a7.bSep07.

ALDH1B1

To date, no human transcriptional variants have been identified for this gene.

ALDH1L1

Five transcriptional variants have been identified for the ALDH1L1 gene (Figure 2, Table 2). The major transcript ALDH1L1_v1 encodes a 902-residue protein, and ALDH1L1_v2 encodes a 912-residue variant. ALDH1L1_v1 and ALDH1L1_v2 differ by an alternative exon 1—resulting in varied translation initiation points on exons 2 and 1 for ALDH1L1_v1 and ALDH1L1_v2, respectively. The ten additional amino acids at the NH₂-terminus of ALDH1L1_v2 are not within any of the three peptide domains previously described for this protein; as such, functional relevance, if any, is unclear. The ALDH1L1_v3 transcript lacks the 151-bp exon 13 present in the other two variants. This represents a significant alteration in the reading frame that introduces an early termination signal and subsequent truncation in peptide translation. This truncation ablates most of the ALDH peptide domain, including its active-site cysteine and glutamate residues; accordingly, ALDH activity for this variant would presumably be null. ALDH1L1_v4 and ALDH1L1_v5 are truncated transcripts with no ALDH peptide domain in either of their resultant translated products.

Figure 2.

Human ALDH1A1 alternatively-spliced variant exon structures. The consensus variant ALDH1A1_v1 is a 13-exon transcript, whereas ALDH1A1_v2 has a shorter sequence due to truncation at its 3’ end. Specifically, ALDH1A1_v2 has a truncated exon 7 and a longer intron 8; its last exon (exon 8) is a 5’ and 3’ truncated subset of the ALDH1A1_v1 exon 9. The translation of ALDH1A1_v2 retains an ALDH peptide domain; however, no active-site residues are readily apparent.

ALDH1L2

The ALDH1L2 gene has three transcriptional variants (Table 2). The major transcript ALDH1L2_v1 encodes a 923-amino-acid protein. ALDH1L2_v2 utilizes an alternate exon 1, a 5’extended derivative of ALDH1L2_v1 exon 13, and lacks exons 1 to 12 of the ALDH1L2_v1 variant. The translation of this variant retains a central portion of ALDH peptide domain but the NH₂-terminal and COOH-terminal formyl transferase peptide domains are ablated. The variant ALDH1L2_v3 lacks the 70-bp exon 1 of the ALDH1L2_v1 variant and encodes an 810-residue protein.

ALDH2 Family

To date, no human transcriptional variants have been identified for this gene.

ALDH3 Family

ALDH3A1

Several alternative splice variants exist within the molecular sequence databases for human ALDH3A1. The consensus gene product is an 11-exon transcript encoding a 50.4-kDa, 453-residue protein. Analysis of cDNA sequences for ALDH3A1 demonstrates a prevalence of three additional variants: ALDH3A1_v2, _v3 and _v4 relative to the ALDH3A1_v1 Reference Sequence (Table 2).

ALDH3A1_v2 comprises only nine exons, but encodes a larger 570-amino-acid variant due to its second exon being a fusion of exon 3, intron 3 and exon 4 (relative to the wild-type ALDH3A1_v1).. ALDH3A1_v3 is also an 11-exon transcript but it differs slightly from the ALDH3A1_v1 transcript by having a 5’ truncation of “GAG” from exon 7 within the coding region.. ALDH3A1_v4 is a 9-exon variant lacking the ALDH3A1_v1 exons 2 and 9. ALDH3A1_v5 is an 8-exon variant resembling ALDH3A1_v2, with regard to the “fusion” exon. However, this variant lacks exon 1 and the “fusion” exon has a 5’ truncation of the 88-bp exon 3 of ALDH3A1_v1. ALDH3A1_v6 is a 10-exon variant lacking the ALDH3A1_v1 exon 7 and truncation of 50 bp from the 5’ portion of exon 8. Lastly, ALDH3A1_v7 is a 10-exon variant lacking the ALDH3A1_v1 exon 2 encoding a functional ALDH peptide domain.

ALDH3A2

Similar to human ALDH3A1, ALDH3A2 has a number of transcriptional variants (Table 2). The primary variant ALDH3A2_v1 is a 10-exon transcript encoding a 485-residue protein expressed in microsomes. ALDH3A2_v2 includes an additional 125-bp exon between exons 9 and 10 (relative to the ALDH3A2_v1 variant), thus encoding a longer protein of 508 amino acids that is expressed in the peroxisomes [29]. The ALDH3A2_v3 and ALDH3A2_v4 variants have coding regions identical to that of the ALDH3A2_v1 and ALDH3A2_v2 variants, respectively, and differ only in exon structure. A number of independent cDNAs within the molecular sequence databases suggest the existence of ALDH3A2_v5—which uses an alternative exon 1 beginning upstream to and including exon 4 of the ALDH3A2_v1 variant.

ALDH3B1

Human ALDH3B1 may have as many as five transcriptional variants, according to the molecular sequence databases for the human ALDH3B1 gene (Table 2). The consensus product is a 10-exon transcript encoding a 468-residue protein. The ALDH3B1_v2 variant lacks exon 3 relative to ALDH3B1_v1; although exon 3 is within the coding region of the peptide, its translation is not associated with the ALDH peptide domain. Therefore, this variant encodes a shorter protein with a complete ALDH peptide domain. The ALDH3B1_v3 transcript has a 3340-bp exon 2—which is a fusion of exon 2, intron 2 and exon 3 of the ALDH3B1_v1 variant. This fusion results in a 3’ shift in the transcript coding sequence and subsequent NH₂-terminal truncation of the peptide. ALDH3B1_v4 lacks exons 1 and 2, plus a 54-base segment from the 5’ end of exon 3 (relative to ALDH3B1_v1) resulting in an NH₂-terminal truncation of the ALDH peptide domain for this protein. ALDH3B1_v5 utilizes a distinct exon 1 and lacks the ALDH3B1_v1 exon 6. There is evidence suggesting a sixth variant, ALDH3B1_v6; the first exon of ALDH3B1_v6 is a 2516-bp fusion of intron 2 and exon 3 of the ALDH3B1_v1 variant and results in an NH₂-terminal truncated protein.

ALDH3B2

Three transcriptional variants have been identified in the sequence databases. ALDH3B2_v1 and ALDH3B2_v2 differ by an alternative exon 1. ALDH3B2_v3 lacks the 100-bp exon 9 present in ALDH3B2_v1, resulting in a shorter protein truncated at the COOH-terminus portion of the ALDH peptide domain.

ALDH4 Family

ALDH4A1_v1 is a 15-exon transcript encoding a 563-amino-acid variant. ALDH4A1_v1 and ALDH4A1_v2 have identical coding regions and subsequently yield identical proteins. The variation between these two transcripts occurs in the last exon (relative to ALDH4A1_v1), because it is transcribed as two separate exons in ALDH4A1_v2: a 154-bp exon 15 and a 359-bp exon 16—both separated by a 1013-bp intron 15, thus yielding a variably sized 3’-UTR. A third variant (described by EMBL-EBI’s Ensembl) lacks the ALDH4A1_v1 exon 4, resulting in a 5’ truncation of the protein’s ALDH peptide domain. ALDH4A1_v4 and ALDH4A1_v5 represent shorter transcripts, yielding peptides truncated at the COOH-terminus with partial ALDH domains and no apparent active site residues (according to Pfam analysis). Another variant, ALDH4A1_v6, has been identified in our laboratory and is being further characterized (W. Black, D. Stagnos, and V. Vasiliou; manuscript in preparation); this transcript lacks exon 12 (relative to ALDH4A1_v1), yet is translated as a splice variant that is missing an internal 51-amino-acid segment.

ALDH5 Family

ALDH5A1_v1 is a 10-exon transcript encoding a 535-amino-acid peptide. ALDH5A1_v2 variant has an additional 39-bp exon transcribed from within intron 4. This exon accounts for 13 additional amino acids within the ALDH peptide domain region of ALDH5A1_v2 (relative to the ALDH5A1_v1 protein). Evidence exists for a third and shorter variant, ALDH5A1_v3, which lacks both 5’ and 3’ exon segments (relative to ALDH5A1_v1). This translates into an NH₂- and COOH-terminal truncated protein that retains a partial ALDH peptide domain, although with no apparent active-site residues.

ALDH6 Family

ALDH6A1_v1 is a 12-exon transcript encoding a 535-amino-acid protein. ALDH6A1_v2 lacks exons 1 through 6 and begins 6-bp upstream from exon 7 (relative to ALDH6A1_v1). The last exon of ALDH6A1_v1 is transcribed as two separate exons in ALDH6A1_v2: a 442-bp exon 6 and a 404-bp exon 7, both separated by a 2237-bp intron. The coding sequence for this transcript ends within exon 6 at the same stop codon as the primary variant, thereby rendering exon 7 irrelevant to the protein’s amino-acid sequence. ALDH6A1_v3 and ALDH6A1_v4 are truncated transcripts at their 3’ ends and comprise exons 1 to 5 and exons 1 to 4 of ALDH6A1_v1, respectively. Both of these variants encode truncated proteins at their COOH-termini; however, they retain a 5’ portion of the ALDH peptide domain.

ALDH7 Family

To date, no human transcriptional variants have been identified for this gene.

ALDH8 Family

Human ALDH8A1 has two transcriptional variants so far identified (Table 2). ALDH8A1_v1 represents the longer transcript encoding a 487-residue protein. ALDH8A1_v2 lacks an in-frame segment within the coding region (exon 6 of ALDH8A1_v1); this translates into a 433-amino-acid splice variant, which has no apparent active-site residues within the ALDH peptide domain.

ALDH9 Family

To date, no human transcriptional variants have been identified for this gene.

ALDH16 Family

Perhaps two transcriptional variants exist for human ALDH16A1 (Table 2). ALDH16A1_v1 is a 17-exon transcript encoding an 802-amino-acid protein. A second variant may be present, although evidence is limited. ALDH16A1_v2 comprises 15 exons. Its exon 6 is a fusion of exon 6, intron 6 and exon 7; its exon 15 is a fusion of exon 16, intron 16 and exon 17 (relative to ALDH16A1_v1). This fusion alters the reading frame of the coding sequence and introduces an early termination codon with subsequent truncation in translation of the peptide.

ALDH18 Family

Alternative splicing of human ALDH18A1 and mouse Aldh18a1 generates two proteins that differ by a 2-amino-acid insertion at the NH₂-terminus of the γ-glutamyl kinase active-site [30]. Exon 6 is 159- and 153-bp in length for ALDH18A1_v1 and ALDH18A1_v2, respectively, yielding the two additional amino acid residues. The shorter variant, ALDH18A1_v2, has high activity in the gut and catalyzes an essential step in arginine biosynthesis. It is inhibited by ornithine, a mechanism by which arginine synthesis can be regulated. The widely expressed longer enzyme ALDH18A1_v1 is necessary for synthesis of proline from glutamate and is insensitive to ornithine inhibition. Impaired function of both the long and short forms, by way of mutations in the human ALDH18A1 gene, may be associated with neurodegeneration, cataracts, and connective tissue diseases [17]. Further studies of these and other ALDH alternative transcripts and protein products will be needed to elucidate their physiological function and significance.

Concluding Remarks

The mammalian ALDH genes identified to date appear to be comprehensive for human, mouse and rat because these genomes are virtually complete. As a result, additional ALDH genes are unlikely to be found in these species, although orthologs and paralogs will continue to be identified in other species as the completion of additional genomes occurs. The human ALDH gene superfamily comprises 19 genes in eleven families and four subfamilies. When compared with the human genome, rat and mouse include an additional gene in the ALDH1A subfamily, namely Aldh1a7. In addition, whereas the human and mouse genomes contain the human ALDH4A1 and mouse Aldh4a1 gene, a rat ortholog has yet to be identified or documented. However, strong evidence for the presence of rat Aldh4a1 exists, located at rat chromosome 5q36. Whereas many mammalian ALDH genes have been identified, several of the protein products encoded by these genes are not yet fully characterized.

Genomic alignment of existing transcript sequences from the molecular sequence databases reveals a number of potential alternatively-spliced transcriptional variants of human, mouse and rat ALDH genes. Yet, little empirical evidence has been reported for these variants in the literature. Further studies will be needed to assess the cell-specific existence of these variants and, ultimately, the functional relevance of such spliced gene products.

Figure 3.

Human ALDH1L1 exon and protein structures for alternatively-spliced transcriptional variants. Most, or all, of the ALDH peptide domain in variants _v3, _v4 and _v5 are ablated and thus ALDH activity is presumed to be nil.