OUP user menu

Evolution of the carboxylate Jen transporters in fungi

Tiziana Lodi, Julie Diffels, André Goffeau, Philippe V. Baret
DOI: http://dx.doi.org/10.1111/j.1567-1364.2007.00245.x 646-656 First published online: 1 August 2007


Synteny analysis is combined with sequence similarity and motif identification to trace the evolution of the putative monocarboxylate (lactate/pyruvate) transporters Jen1p and the dicarboxylate (succinate/fumarate/malate) transporters Jen2p in Hemiascomycetes yeasts and Euascomycetes fungi. It is concluded that a precursor form of Jen1p, named here preJen1p, arose by the duplication of an ancestral Jen2p, during the speciation of Yarrowia lipolytica, which was transferred into a new syntenic context. The Jen1p transporters differentiated from preJen1p in Kluyveromyces lactis, before the Whole Genome Duplication (WGD), and are conserved as a single copy in the Saccharomyces species. In contrast, the ancestral Jen2p was definitively lost just prior to the WGD and is absent in Saccharomyces.

  • carboxylate transporters
  • yeast
  • evolution
  • synteny


The Kluyveromyces lactis KlJEN1 and KlJEN2 genes encode yeast plasma membrane transporters belonging to the facilitator superfamily (MFS) (André, 1995,; Nelissen, 1995). KlJen1p, first identified in Saccharomyces cerevisiae (Casal, 1999; Akita, 2000; ; Lodi, 2002), is a lactate/pyruvate symporter (TC 2.A.1.12.2) that belongs to the Sialate:H+Symporter (SHS) family (Saier, 2000). KlJen2p, which shows 52% amino acid similarity to KlJen1p, is a succinate/fumarate/malate transporter recently identified in K. lactis but absent in Sac. cerevisiae (Lodi, 2004; Queirós, 2007).

The Sac. cerevisiae genome was completely sequenced and annotated 10 years ago (Goffeau, 1996). Over the past few years, the nearly completed genome sequence of over a dozen other yeast species became available (Souciet, 2000; Cliften, 2003; Kellis, 2003, 2004; Dietrich, 2004; Dujon, 2004). Most of the sequenced yeast species are members of the Hemiascomycetes subdivision of fungi, to which budding yeasts belong. Several other nearly completed genome sequences from Euascomycetes, such as Neurospora crassa (Galagan, 2003), and the Aspergillus species (Galagan, 2005; Machida, 2005; Nierman, 2005) or from Archeascomycetes, such as Schizosaccharomyces pombe (Wood, 2002), also became available. This large panel of fungi species offers a unique opportunity to compare a homogeneous phylogenetic group of eukaryotic organisms and to examine specific questions in terms of genome evolution (Dujon, 2006).

In this perspective, the Yeast Gene Order Browser (YGOB) is a novel tool for the analysis of synteny relationships in hemiascomycetous genomes preceding or following the Whole Genome Duplication (WGD) that occurred between the Kluyveromyces and Saccharomyces speciation (Byrne & Wolfe, 2006). The synteny approach has recently been further developed for functional annotation and evolution tracing of protein families in the Hemiascomycetes (Diffels et al., in preparation).

In this paper 35 proteins encoded by putative orthologs of the KlJEN1 and KlJEN2 genes in 17 different fungi, of which 13 are Hemiascomycetes and four are Euascomycetes, are analyzed. According to the nomenclature used until now (Casal, 1999; Lodi, 2004), they are called Jenp here.

Phylogenetic, blast and synteny analyses have been combined to distinguish and trace the evolution of the Jenp clusters in these fungi. Moreover, specific Jenp amino acid sequence motifs were traced in each cluster. An evolution pattern is proposed for the gain and loss of the phylogenetic Jenp clusters among Hemiascomycetes and Euascomycetes.

Materials and methods

Transporter selection

The following genomes were screened to highlight the putative orthologs of KlJEN1 and KlJEN2 using blast (Altschul, 1997): Sac. cerevisiae, Sac. paradoxus, Sac. mikatae, Sac. kudriavzevii, Sac. bayanus, Sac. castellii, Candida glabrata, Zygosaccharomyces rouxii, Sac. kluyveri, Kluyveromyces thermotolerans, K. waltii, K. lactis, Ashbya gossypii, C. albicans, Debaryomyces hanseni, Yarrowia lipolytica, Neurospora crassa, Aspergillus nidulans, Asp. fumigatus, Asp. oryzae and Schizosaccharomyces pombe. The blast search was carried out using the ScJEN1 coding sequence as a query on the available databases: SGD (http://www.yeastgenome.org/), Genolevures (http://cbi.labri.fr/Genolevures/blast/index.php), Duke (http://ashbya.genome.duke.edu/), CADRE (http://www.cadre.man.ac.uk/), MNCDB (http://mips.gsf.de/genre/proj/ncrassa/), CandidaDB (http://genolist.pasteur.fr/CandidaDB/) and Schizosaccharomyces pombe GeneDB (http://www.genedb.org/genedb/pombe/). BLAST STATION was also used locally. The protein data from Sac. paradoxus, Sac. mikatae, Sac. kudriavzevii, Sac. bayanus, and Sac. kluyveri were downloaded from SGD (ftp://genome-ftp.stanford.edu/pub/yeast/data_download/sequence/fungal_genomes/) whereas those of Z. rouxii, K. thermotolerans and K. waltii were provided by the Genolevures consortium. The best hits above the first discontinuity in the blastE-value distribution were retained. Their E-value was always lower than 10−65.

A ‘working name’ was introduced to describe the different Jen proteins (Table 1). The first four letters of this name indicate the yeast genus and species; an additional letter indicates the chromosome on which the gene is located (when unknown, the chromosome number is replaced by X). J1/J2 indicates whether the transporter is a member of the Jen1 or the Jen2 group (J0 is used when the attribution is dubious). The final numbers 01–04 indicate different paralogs present in the same organisms. The last two characters of the working name SAPA-X1 correspond to contig c445 of Sac. paradoxus, SAMI-X1 to contig 1028 of Sac. mikatae, SAKU-X1 to contig 1994 of Sac. kudriavzevii, SABA-X1 to contig 579 of Sac. bayanus, KLWA-X1 to contig 163 of K. waltii, SAKL-X1 to contig 2083 of Sac. kluyveri, SAKL-X2 to contig 2253 of Sac. kluyveri, SAKL-X3 to contig 1959 of Sac. kluyveri, NECR-X1 to contig 3nc270 of N. crassa, and NECR-X2 to contig 9a2 of N. crassac.

View this table:

Jen transporters identified in fungi

Identification Numbers
SpeciesCDS NameWorking NameEMBLSwiss-ProtLength (aa)TMS
K. lactisKlJEN1(KLLA0E16313g)KLLA-E-KLJEN1AJ585426Q70DJ760012
KlJEN2 (KLLA0F10043g)KLLA-F-KLJEN2AJ627630Q701Q952811
Sac. cerevisiaeJEN1 (YKL217w)SACE-K-JEN1U24155P3603561612
Sac. paradoxusSpar_Contigc445_ORFP12938SAPA-X1-J10161612
Sac. mikataeSmik_Contig1028_ORFP12909SAMI-X1-J10161612
Sac. kudriavzeviiSkud_Contig1994.4SAKU-X1-J10161410
Sac. bayanusSbay_Contig579.2SABA-X1-J10161512
K. thermotoleransKLTH-IPF6111KLTH-G-J20154211
K. waltiiKLWA-IPF7204KLWA-X1-J20153411
Sac. kluyveriSklu_Contig2083.1SAKL-X1-J10159810
A. gossypiiAAR192CpASGO-A-J001AAS50559Q75E8850010
D. hanseniiDEHA0D20427gDEHA-D-J202CAG87482Q6BR6251110
C. albicansCA5478CAAL-D-J001EAK98054Q5A5U251310
Y. lipolyticaYALI0B19470gYALI-B-J101CAG83351Q6CE1452910
N. crassancu07607.2NECR-X1-J201EAA33605Q7SB4755711
A. fumigatusAfu7g05550ASFU-G-J202EAL86916Q4WGM550111
A. nidulansAN6703.3ASNI-A-J20149510
A. oryzaeAO090005000420ASOR-A-J20150811
  • * Predicted by the tmhmm program.

Phylogenetic tree

The multiple alignments of putative Jen transporters were carried out by the t-coffee program (Notredame, 2000). The phylip suite (Felsenstein, 1989) was used for phylogenetic tree construction. The evolutionary distances were estimated by protdist and the clustering by the neighbour programs. The unrooted tree was drawn using treeview (Page, 1996).

Synteny analysis

Synteny analysis was performed as reported by Diffels et al. (in preparation) on the amino acid sequences of five neighbour genes located on the left and on the right of each JEN gene. A neighbour is defined as a predicted amino acid sequence with a length of over 100 amino acids that is located beside the JEN encoded transporter.

A systematic blast comparison between all neighbour sequences was carried out. Sequences below a threshold E-value lower than 10−10 were collated and called ‘homologous neighbours’. Then, all the blocks of genes, constituted by the JEN encoded transporter and its neighbour-translated coding sequences (CDS), were compared to each other according to the presumed evolution course of the analysed species. Two blocks containing at least one homologous neighbour were considered as sharing a synteny relationship and thus belonging to the same evolution event. In this analysis, it is postulated that only the proteins with similar amino acid sequences and with the same chromosomal environment are real orthologs.

Motif search

All the protein sequences were submitted to the tmhmm program (Krogh, 2001) to locate the transmembrane spanning helices. The transporters were aligned with t-coffee (Notredame, 2000). The multiple alignments allowed one to detect the presence of amino acid sequence gaps and eventually sequence motifs distinguishing the Jenp transporters. The meme program (Bailey & Elkan, 1994) was used to confirm the motifs detected by visual inspection.


Identification of Jen transporters in fungi

SGD, Genolevures and other yeast genomes databases were searched for JEN sequence homologs by blast analysis using ScJEN1 coding sequence as query. Table 1 lists the 35 homologous sequences found in 13 different Hemiascomycetes and four Euascomycetes genomes. Table 1 also shows that the length of the Jenp amino acid sequences varies from a minimum of 478 in Jen2p from A. gossypii (ASGO-B-J201) to a maximum of 616 residues in Jen1p from Sac. cerevisiae, Sac. paradoxus and Sac. mikatae. All Jenp displayed transmembrane structures typical of secondary transporters, with 9–12 predicted transmembrane spanning helices (TMS). This indicates that KlJen1p and KlJen2p belong to a protein family conserved through the evolution from Euascomycetes to Hemiascomytes. In the different yeast species, the number of JEN paralogs varied from one (as in Sac. cerevisiae and in all Saccharomyces sensu stricto) to six (as in Yarrowia lipolytica). In C. glabrata, Sac. castellii, Z. rouxii and Sch. pombe, no Jenp were found. In Sch. pombe the presence of the malic acid transporter Mae1p could justify the lack of Jenp, but blast analysis does not highlight Mae1p in the three other species. As C. glabrata, Sac. castellii and Z. rouxii have not yet been studied for controlled growth on mono- or dicarboxylate, one cannot rule out that other carboxylate transporters undetected at this time could play a role in these species. As C. glabrata grows in the blood, the absence of carboxylate transporters could be explained. Moreover, the genome sequences are currently still incomplete and some transporters could have been missed.

blast and phylogenetic analyses of yeast Jen transporters

In several cases, blast analysis did not suffice to predict the phylogenetic group (Jen1 or Jen2) to which the Jen transporters belong (Table 2). For example, DEHA-F-J101 and DEHA-F-J201 display sequence similarity to both KlJen1p and KlJen2p.

View this table:

blast analysis of yeast Jen transporters

Working NameCDS NameKLJEN1KLJEN2Origin

From the sequence similarity data, a phylogenetic tree was drawn (Fig. 1). Two distinct patterns were observed. On the one hand, most of the hemiascomycetous transporters cluster either around the functional model KlJen1 (cluster A in Fig. 1) or around KlJen2 (cluster C). On the other hand, transporters cluster according to a systematic appurtenance; Y. lipolytica members in cluster D and all Euascomycetes members in cluster E. Three Jen transporters, CAAL-C-J101, DEHA-F-J101 and DEHA-FJ201, do not fit in any other cluster and constitute the cluster B.


Phylogenetic tree of the Jen-like transporters.

Synteny analysis

Chromosomal regions are considered syntenic, when adjacent genes are present in a conserved order in two (or more) genomes. In closely related species, large syntenic blocks are observed. The size of the syntenic blocks decreases when the phylogenetic distance increases.

The results of synteny analysis are summarized in Figs 2 and 3 for Jen1p and Jen2p, respectively. In both figures, boxes represent individual gene products encoded in the chromosomal stretches surrounding JEN1 or JEN2 in different fungal genomes. The earliest divergent genomes are set at the bottom of the figures and the most recent species on the top. The Jen1 or Jen2 proteins are represented in red. The gene products of neighbour genes are identified by a code name similar to that used for the Jen names (supplementary Tables S1 and S2). The orthologous families are distinguished by specific colours and by the same first letter in front of the working names. When no syntenic information exists (e.g. when CDS are localized in a telomeric region or when small contigs are not assembled) no boxes are represented. Sometimes, when the synteny relationship between two blocks of close species could not be directly established, i.e. when the blocks share no common homologous neighbours (see Fig. 2-KLLA/Saccharomyces), the analysis of a third block in another species (see Fig. 2-YALI, chr B) allows one to establish indirectly the relationship between the three species.


Syntenic context of the Jen1-like transporters.


Syntenic context of the Jen2-like transporters.

Figure 2 shows that a continuous syntenic context can be drawn for Jen1p from Y. lipolytica to Sac. cerevisiae. Figure 3 shows context continuity from Aspergillus to K. thermotolerans for Jen2p.

Table 3 lists the members of Jen1 or Jen2 clusters obtained by combination of synteny and blast analyses. Such analyses clearly allocate each of them to the Jen1p or Jen2p cluster, except for the following four undetermined cases: CAAL-D-J001, ASGO-A-J001, SAKL-X3-J001 and ASFU-H-J001, for which no synteny could be detected and which were consequently named Jen0p. These undetermined Jen0p cases have a higher sequence similarity to Jen2p than to Jen1p.

View this table:

Members from Jen clusters in different yeast species

SpeciesJEN1 ContextJEN2 ContextJEN0 ContextTotal

Search of molecular motifs specific to Jen1p and Jen2p

Presence/absence of TMS11

The TMS profiles of KlJen1p and KlJen2p in Fig. 4 illustrate that in the Jen2 transporters the penultimate C-terminal helix (that named TMS11 here) is predicted with a probability lower that 20%. In contrast, in the transporters classified as member of the Jen1 cluster, the existence of TMS11 is predicted with high probability. As shown in Table 4, the following exceptions were noticed: (i) TMS11 is predicted with high probability in all three D. hansenii Jenp, one of which was assigned to the Jen1p cluster and two of which were assigned to the Jen2p cluster; (ii) TMS11 is predicted with low probability in YALI-B-J101, YALI-C-J102 and ASGO-F-J101, which were assigned to the Jen1p cluster.


Prediction by tmhmm of the transmembrane spans of KlJen1 and KlJen2.

View this table:

Protein sequence motifs in Jen putative transporters

Working NameJEN1JEN2TMS 11ProlineTMS6/7Gap
  • * Member of Jen1p cluster (1) or not (0).

  • Member of Jen2p cluster (1) or not (0).

  • TMS11 predicted with high probability (1) or not (0).

  • § Presence of proline near the C-terminal (1) or substitution with Histidine (0).

  • Gap between the TMS6 and TMS7 (1), no deletion (0).

Proline/histidine substitution in the C-terminal regions

Analysis of the amino-acid sequences of Jen1p and Jen2p reveals a striking substitution of a proline residue in Jen2p by a histidine residue in Jen1p located near C-terminal end (see Fig. S1). The residues are localized in a stretch that displays high conservation. Such substitution may have drastic structural and functional consequences, as proline is known to break transmembrane α-helixes while histidine is often involved in acid–base catalysis (Lehninger, 1982). There are five exceptions to this rule, as shown in Table 4. Indeed YALI-B-J101, YALI-C-J102, DEHA-F-J101, CAAL-C-J101 and ASGO-F-J101 belong to the syntenic Jen1p context but nevertheless show the Jen2p proline residue.

Presence/absence of a sequence deletion between TMS6 and TMS7

It was also observed that the distance between the sixth and seventh TMS helixes is shorter in KlJen2p than in KlJen1p (Fig. 4, middle). Figure 5 shows that this correlates with a 7–15 amino acid deletion (gap) in the KlJen2p amino acid sequence. This gap between TMS6 and TMS7 is present in all the members of Jen2p cluster. Also, its presence correlates with the presence of the Jen2p specific proline near the C-terminal and thus the same exceptions were found for the five Jen1p members exhibiting the proline marker of Jen2p (Table 4).


Alignment of KlJen1 and Jen2 amino acids sequences showing the sequence deletion (gap) in the intracytosolic loop between TM6 and TM7 from Jen2 transporters. Translated CDS names are indicated on the left. The numbers on the right indicate the position in the amino acid sequence.


For 31 out of the 35 identified Jenp members, the synteny analysis distinguishes clearly the Jen1p from the Jen2p filiations, as summarized in Fig. 6. Figure 6 also shows that the ancestor of the Jenp subfamily is an ortholog of Jen2p, which can be detected early in Euascomycetes species such as Neurospora and Aspergillus. The Jen2p orthologs are characterized by synteny as well as by blast sequence similarity. The Jen2p members contain three specific motifs not shared by Jen1p: TMS11 predicted with low probability, proline near the C-terminal, and a sequence gap between TMS6 and TMS7. The only exceptions are the two D. hanseni transporters attributed to the Jen2 cluster by synteny in which the TMS11 is predicted with high probability.


Summary of similarity, synteny analysis and motif identification of the members of fungal Jenp clusters. Circles represent different Jen transporters, classified as Jen1 (in grey), Jen2 (in orange), Jen0 (in white). Boxes represent motifs: the first box on the left is related to TMS11 (green box if TMS11 is predicted with high probability, red box if it is predicted with low probability); the central box is related to the presence/absence of proline near the C-terminal of the proteins (green box if proline is present, red box if proline is substituted by histidine); the third box is related to the gap between the TMS6 and TMS7 (green box if there is a gap, red box if the gap is not present). The level of amino acid sequence divergence between Jen1p and Jen2p is related to the horizontal distance between Jen2p and the Jen1p (or preJen1p) circles (highly similar sequences are nearby). The WGD is indicated by the horizontal line between K. thermotolerans (KLTH) and Sac. bayanus (SABA). The Euascomyces and Hemiascomycetes are seperated by another horizontal line between N. crassa (NECR) and Yarrowia lipolytica (YALI).

The Jen2p are conserved through the Hemiascomycetes phylum from Y. lipolytica up to K. thermotolerans but are absent in the post-WGD species such as the Saccharomyces sensu largo and sensu stricto species. Thus, it is highly probable that the loss of Jen2p happened just prior to the WGD. It is therefore assumed that the dicarboxylate transport function (succinate/fumarate/malate) of Jen2p may be conserved from the filamentous fungi to the Kluyveromyces yeasts.

On the other hand, the monocarboxylate (lactate, pyruvate) transporter Jen1 is conserved from K. lactis through all Saccharomyces species. Jen1p can be distinguished from Jen2p by synteny, similarity and motifs. KlJen1p did not arise by duplication of KlJen2p and its emergence is not related to the WGD. Jen1p seems to have arisen from an ‘intermediate’ carboxylate transporter which can be traced as a duplication of Jen2p in Y. lipolytica. Indeed in Y. lipolytica Jen2p is highly amplified and produces not only four Jen2 paralogs but also two members which, by synteny analysis, belong to Jen1p. However, by other criteria such as sequence similarity and protein motifs these transporters are similar to Jen2p. Similar ‘intermediate’ members of the Jenp subfamily are present in D. hansenii, C. albicans and A. gossypii. They differ from Jen1p by at least one of the three Jen1p-specific sequence motifs. It is only in K. lactis that a typical Jen1p emerges, while the ‘intermediate’ form is lost. The YALI-B-J101, YALI-C-J102, DEHA- F-J101, CAAL-C-J101 and ASGO-F-J101 ‘intermediate’ variants as considered as precursor forms of Jen1p that are therefore named ‘preJen1’ in Table 4 and Fig. 6.

As a practical consequence of this analysis, it is suggested that in addition to the 2.A.1.12.2 cluster already allocated to Jen1p in the transport classification (TC) (Saier, 2000), the new TC numbers 2.A.1.12.3 be given to the Jen2p cluster and 2.A.1.12.4 to the novel ‘preJen1p’ cluster within the bacterial and fungal SHS subfamily. However, the exact chemical nature of the substrates transported by members of the preJen1p cluster remains to be elucidated.

In brief, the combination of synteny analysis, sequence similarity and motif analysis allowed the tracing of evolution of the fungal Jenp subfamily. It is worth noting that such archeogenomological reconstitution would have been wrongly deduced from the sole blast analysis or from the sole analysis of the phylogenetic tree, which are not sufficiently discriminating and may lead to artificial clustering of members from the distant Y. lipolytica and Euascomycetes species, as illustrated in Fig. 1.

Author contribution

T.L. and J.D. contributed equally to this work.

Supplementary material

Table S1. Identification of JEN1 homologous neighbour.

Table S2. Identification of JEN2 homologous neighbour.

Figure S1. Near C-terminal stretches of amino acid sequence of Jenp transporters.


This work was funded by grants from FIL Program, University of Parma. We thank all our colleagues of the Genolevures consortium. A special thank to Marie-Line Seret for helpful discussion. J.D. was funded by the FSR program, Université catholique de Louvain and by a FRIA fellowship, Fonds national de la Recherche scientifique, Belgique.


View Abstract