Biochemical and Biophysical Research Communications 336 (2005) 175–189

BBRC
www.elsevier.com/locate/ybbrc

The repertoire of solute carriers of family 6: Identification

of new human and rodent genes
Pär J. Höglund, Dijana Adzic, Sara J. Scicluna, Jonas Lindblom, Robert Fredriksson *
Department of Neuroscience, Uppsala University, BMC, Box 593, 751 24 Uppsala, Sweden

Received 3 August 2005

Available online 18 August 2005

Abstract

Tremendous amount of primary sequence information has been made available from the genome sequencing projects, although a
complete annotation and identification of all genes is still far from being complete. Here, we present the identification of two new
human genes from the pharmacologically important family of transporter proteins, solute carriers family 6 (SLC6). These were
named SLC6A17 and SLC6A18 by HUGO. The human repertoire of SLC6 proteins now consists of 19 functional members and
four pseudogenes. We also identified the corresponding orthologues and additional genes from mouse and rat genomes. Detailed
phylogenetic analysis of the entire family of SLC6 proteins in mammals shows that this family can be divided into four subgroups.
We used Hidden Markov Models for these subgroups and identified in total 430 unique SLC6 proteins from 10 animal, one plant,
two fungi, and 196 bacterial genomes. It is evident that SLC6 proteins are present in both animals and bacteria, and that three of the
four subfamilies of mammalian SLC6 proteins are present in Caenorhabditis elegans, showing that these subfamilies are evolutionary
very ancient. Moreover, we performed tissue localization studies on the entire family of SLC6 proteins on a panel of 15 rat tissues
and further, the expression of three of the new genes was studied using quantitative real-time PCR showing expression in multiple
central and peripheral tissues. This paper presents an overall overview of the gene repertoire of the SLC6 gene family and its expres-
sion profile in rats.
Ó 2005 Elsevier Inc. All rights reserved.

Keywords: Genome projects; Profile hidden Markov models; SLC6; Evolution

Genes coding for membrane proteins are one of the carriers [9] and one of these is the solute carrier family 6
largest groups of genes in vertebrate genomes. Mem- (SLC6). To separate the letter 6 of the solute carrier type
brane proteins are claimed to constitute more than from the family member number a capital A is intro-
10% of all genes in the human [1] and mouse genomes duced, i.e., SLC6A4 (solute carrier family 6) separates
[2], respectively. Important subgroups of membrane the gene family name ‘‘6’’ from the family member num-
proteins are G-protein coupled receptors [3,4], single ber ‘‘4.’’ The SLC6 family of proteins acts as specific
pass tyrosine kinase receptors [5,6], ion channels [7], transporters for neurotransmitters, amino acids, and
and transporters or solute carriers. Solute carriers con- osmolytes like betaine, taurine, and creatine. The neuro-
trol the uptake and flow of various substances such as transmitters whose levels are controlled by SLC6
sugar, amino acids, nucleotides, inorganic ions, and proteins are low molecular weight molecules like acetyl-
drugs over the cell membrane, either passively or active- choline, dopamine, serotonin, norepinephrine, and ami-
ly [8]. There are currently 43 different families of solute no acids like c-aminobutyric acid (GABA), glycine, and
aspartate. The SLC6 proteins are Na+ cotransporters
and the energy for the transport of the solute against
*
Corresponding author. Fax: +46 18 51 15 40. its concentration gradient is provided by the electro-
E-mail address: [email protected] (R. Fredriksson). chemical gradient for sodium ions. In some cases, such

0006-291X/$ - see front matter Ó 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.bbrc.2005.08.048
176 P.J. Höglund et al. / Biochemical and Biophysical Research Communications 336 (2005) 175–189

as for the SLC6A1 (GAT1), chloride ions are also There has been confusion regarding the nomencla-
cotransported, although this is variable between the ture, naming of orthologous proteins in different species,
members of the family. In addition, the serotonin trans- and the repertoire of functional proteins for the SLC6A
porter SLC6A4 requires the counter transport of potas- family of proteins. For example, the GABA neurotrans-
sium ions but the role of chloride or potassium ions in mitter transporter 2 (SLC6A13) in human is ortholo-
this transport process is not well understood. gous to the GABA 3 transporter in mouse. The
SLC6 family members are also known to be impor- human creatine transporter type 2 (SLC6A10) has been
tant for a number of pathological conditions and several localized to 16p11.2 by two studies [19,20]. This paper
of them are potential drug targets that are pursued by and one other study have identified the SLC6A10 gene
the pharmaceutical industry. Mutations in the SLC6A2 as a pseudogene with an early stop codon [21]. However,
(norepinephrine transporter) cause orthostatic intoler- this gene has been considered to be a functional gene
ance. Orthostatic intolerance is a syndrome character- both in Unigene at NCBI as well as in a recent review
ized by lightheadedness, fatigue, and syncope, and is article [8]. It is known that the family SLC6 is ancient
associated with postural tachycardia and plasma norepi- and exists in eukaryotes as well as in eubacteria and
nephrine concentrations that are disproportionately archaebacteria [22]. However, evolutionary studies of
high in relation to sympathetic outflow. The dopamine the SLC6 family are scarce. In 1998, it was concluded
transporter SLC6A3 (DAT1) that acts to take released that this gene family existed in insects, bacteria, and
dopamine back up into presynaptic terminals has been Caenorhabditis elegans but not in yeast and fungi [22].
implicated in human disorders such as parkinsonism, Tremendous amount of primary sequence information
Tourette syndrome, and substance abuse. The serotonin has become available from recent sequencing projects,
transporter encoded by SLC6A4 is the target of an providing a near to full coverage of the entire genomes
important class of antidepressant drugs, the serotonin from a diversity of animal, plant, and bacteria species,
selective reuptake inhibitors. It has been shown to take although a complete annotation and identification of
part in numerous clinical conditions such as autism, all gene is still far from complete [23] and new genes
depression, neuroticism, and obsessive compulsive dis- are continuously being described [24–26]. The complete
order. There is a high prevalence of SLC6A8 mutations genome material makes comprehensive analysis of the
in X-linked mental retardation. Common features of gene repertoire in multiple genomes possible.
X-linked mental retardation are neurological distur- In this study, we aim to provide a complete view of
bances including seizures, behavioral problems, speech the gene repertoire of the SLC6 family of proteins. We
delay, and inability to engage in structured play [10]. accomplished this through searches in the human,
Mutations in the predominantly neutral amino acid mouse, and rat genomes using various bioinformatic
transporter SLC6A19 are the cause of Hartnup disorder. methods. Several new members of the SLC6 family were
The SLC6A14 gene encodes an amino acid transporter, found and we present genomic structures, orthologous
which potentially regulates tryptophan availability for relationships, phylogenetic grouping, and EST expres-
serotonin synthesis and thus possibly affects appetite con- sion charts for the entire SLC6 family. Moreover, we
trol and mood. Recently, a Finnish–Swedish study has conducted an evolutionary study on the SLC6 repertoire
shown that polymorphism in the SLC6A14 gene is asso- in 13 different species. We also performed reverse trans-
ciated with obesity. A replication study with obese criptase PCR on a tissue panel of rat mRNA to provide
French subjects reconfirmed this conclusion [11,12]. a quantitative expression profile of the entire SLC6
Cloning of the human transporters and their rodent family.
orthologues; SLC6A1 (GABA), (SLC6A2) (norepineph-
rine/dopamine), SLC6A3 (dopamine), and SLC6A4
(serotonin) in 1990–1994 marked the beginning of the Results
identification of genes in the SLC6 family [13–16]. Addi-
tional genes that code for monoamine and amino acid Our initial strategy was to download all known mem-
carriers were identified during the late 1990s. Still, several bers of the human, mouse, and rat SLC6 family. Known
additional genes belonging to this family are ‘‘orphans’’ human SLC6 sequences were identified on the HUGO
or without known functions [17,18]. The common Gene Nomenclature Committee Homepage as well as
structural feature of SLC6 transporters is the presence from GenBank (http://www.ncbi.nih.gov/Genbank/).
of a putative 12 transmembrane (TM) region with the Known mouse sequences were identified on the Mouse
N- and C-terminal regions located inside the plasma Genome Informatics web page and from GenBank.
membrane. The transporters are all relatively large and The rat genome database RatMap contained only six
complex proteins, and consist of between 590 and 800 SLC6 sequences and therefore we also searched Rat
amino acid residues. SLC6 proteins have not yet been Genome Database (http://rgd.mcw.edu/). The protein
crystallized and hence the precise structure of the SLC6 sequences were thereafter downloaded from the Entrez
neurotransmitter is not very well determined. web page (http://www.ncbi.nlm.nih.gov/entrez/). The
 P.J. Höglund et al. / Biochemical and Biophysical Research Communications 336 (2005) 175–189 177

known transporters retrieved were the human SLC6A1- and SLC6A18 to adhere to the accepted nomenclature.
SLCA16 and SLC6A20, the mouse transporters SLC6A18 was present in the sequenced part (21,243
mSLC6A1-A9, mSLC6A10-15 as well as mSLC6A20, cDNA clones) of a library in a large-scale cDNA
and the rat transporters rSLC6A1-A9, rSLC6A10-15, sequencing project and automatically annotated as
and rSLC6A18. The previously unknown gene Xtrp2 [30], but was not characterized further. We also
SLC6A19 was released and added to our data set during found one mouse gene and three rat genes, and these
the finalization of the study [27,28]. were named mSLC6A19, and rSLC6A14, rSLC6A19,
To investigate whether additional SLC6 genes exist and rSLC6A20 according to the orthologous human
in human and rodents, a profile HMM was download- SLC6 family member. The repertoire in human also
ed from http://www.sanger.ac.uk/cgi-bin/Pfam/getacc? contains two previously reported pseudogenes and dur-
PF00209. This profile was used to identify SLC6 like ing our investigation one additional pseudogene,
sequences using HMMsearch from the 2.3.2 release of SLC6A10pB, was identified. The gene repertoire in hu-
the HMMER package [29]. We also used BLAST- man, mouse, and rat consists of 19 genes in each species
searches with known SLC6 proteins as baits. The and that mouse and rat displays an identical 1:1 phylo-
searches were performed in various databases as speci- genetic relationship. On the contrary, the repertoire in
fied under Materials and methods. This resulted in the rodents and human is not identical as SLC6A16 does
identification of two new human genes. We approached not exist in mouse and rat while the X transporter pro-
the HUGO Genome Nomenclature committee and these tein 3 similar 1 (Xtrp3s1) does exist not in human (see
were confirmed to be novel and were named SLC6A17 Fig. 1).

Fig. 1. A phylogenetic analysis of the solute carriers family 6 (SLC6) in human, mouse and rats. Maximum likelihood with 100 replicas was used to
calculate phylogenetic trees. The family groups into five subgroups: I, GABA; II, Monamine; III, Amino Acid; and IV, Orphans.
178 P.J. Höglund et al. / Biochemical and Biophysical Research Communications 336 (2005) 175–189

The new genes identified from the HMM-searches the human genome using BLAT at the UCSC website,
were found as GENSCAN gene predictions. GEN- we found that there are two adjacent pseudogenes about
SCAN and other ab initio gene prediction programs 890 kb apart, SLC6A10pA (32797531–32799840) and
are known to detect around 80% of all exons, although SLC6A10pB (33690486–33692794). We downloaded
the exact exon–intron boundaries are almost newer cor- these genome sequences from UCSC, produced align-
rectly predicted. Therefore, before phylogenetic analysis, ment between these genomic pieces with ClustalW, and
the new genes have to be manually assembled using used Megalign to calculate a nucleotide percentage iden-
other nongenomic information. Below we show the ori- tity. We saw that these two pseudogenes have a 99.6%
gin and how each protein was assembled. identity. SLC6A8 has a 96.8 and 97.0 percentage nucle-
otide identity to SLC6A10p (32797531–32799840) and
Human SLC6A10p (33690486–33692794), respectively.

SLC6A17 (DJ1003J2_NTT4) was found through Mouse

HMM-searches. It had a total length of 436 amino
acids, which is shorter than other members of the family. Mouse SLC6A16 (XP_145587) was derived from a
We extended the sequence by using the previously GNOMON gene prediction, which was further support-
known 724 amino acid long mouse ortholog ed by EST evidence. Our searches yielded 11 ESTs and 1
(NP_758475) as bait in BLAT searches against the cDNA sequence, and all of these were expressed in tes-
human genome. We extracted the corresponding human tis. 94 out of 728 amino acids, including three splice sites
DNA sequence and used it to search for human and were not covered by EST or cDNA. BLAT searches re-
mouse ESTs and cDNA. We used 18 human EST vealed that there is a local duplication of SLC6A16 in
sequences to correct splice site predictions. There were both mouse and rat. The genes are located at 7qB2, only
still gaps in the protein and therefore we added five 5 kb apart. These recently duplicated genes in both
mouse cDNA and 11 mouse ESTs, which resulted in mouse and rat are most likely pseudogenes and were
the entire sequence of human SLC6A17 being verified. therefore excluded from our analysis.
The human SLC6A17 has a 97% and 97.6% identity to mSLC6A19 has one cDNA and two EST sequences
mouse and rat, respectively. that cover the whole protein. During the writing process,
SLC6A18 was found through searches with BLAST another group cloned this protein and linked it to the
using SLC6A16 as bait. The identified sequence was Hartnup disease [31].
gi|34785074 (FLJ31236). In addition, 1 human cDNA
and 7 human EST sequences were assembled, which Rat
covered the whole sequence. The SLC6A18 has 82.4%
and 82.5% amino acid identity to the mouse and rat The rSLC6A14 (XP_233305) was inspected using
SLC6A18, respectively. alignments with the human and mouse orthologues of
SLC6A21p (Hs19_11347_28_95_2) was found when SLC6A14. Six rat and 18 mouse ESTs aligned. The se-
searching with the HMM model against the GENSCAN quence and the exon–intron-boundaries were verified
assembly 28. It had a length of 420 amino acids and is to be correctly predicted.
located at 19q13.33 next to SLC6A16. The length of The rSLC6A19 was derived using a GNOMON gene
known SLC6 proteins ranges from 591 (rSLC6A20) to prediction. In the original prediction, it was 718 amino
799 amino acids (rSLC6A5). The consensus sequence acids with 14 exons. It was corrected using three rat
for the PFAM00209 (sodium: neurotransmitter sym- EST, two mouse cDNA, and 38 mouse EST sequences.
porter family) model has a length of 536 amino acids. The final sequence is a 634 amino acid protein with 12
Searches in all species for ESTs and mRNA yielded no exons. The exon–intron boundaries were corrected by
hits. The short length and the absence of any mRNA the cDNA, mRNA, and mouse, human alignment.
or EST make it likely that SLC6A21 is a pseudogene. The rat SLC6A16 was found by searching the pre-
SLC6A10p is a recent duplicate of the creatinin trans- dicted GNOMON rat dataset based on the NCBI build
porter SLC6A8 protein that was identified by Iyer et al. 2. There is a local duplication of the SLC6A16 pseudo-
[19]. It is located at 16p11.2 and it is concluded that gene. The pseudogenes are located on 1q22 and are 9 kb
there was duplication between 16p11.1 and Xq28. apart. The use of RT-PCR revealed that the rSLC6A16
SLC6A8 is located at Xq28. The duplication probably is not expressed in our rat tissue panel. The lack of
occurred within recent evolutionary time (7–10 mya) expression supports the conclusion that the SLC6A16
[21]. Predicted translations of exons and RT-PCR anal- duplicate is a pseudogene.
yses revealed that this paralog to SLC6A10 on chromo- The rSLC6A20 (XP_217302.2) was corrected by
some 16 probably is a pseudogene as there is a one base alignment with the human and mouse orthologue as well
substitution in exon 4 from Trp (TGG) to a stop codon as rat ESTs. Searching with BLAT shows that the gene
(TGA) [21]. While searching the May 2004 assembly of is located adjacent to rXtrp3s (Q64093). The relation-
 P.J. Höglund et al. / Biochemical and Biophysical Research Communications 336 (2005) 175–189 179

ship is the same in mouse where rXtrp3s is located next and in the CNS. The SLC6A8 is also very highly ex-
to mSLC6A20. pressed in tumor tissues (102 ESTs). Although there
are numerous cases in Table 2 where the expression pat-
Phylogeny tern is similar between human and mouse, there are a
surprisingly high number of cases where the expression
Phylogenetic analysis with maximum likelihood, pattern differs remarkably. Some of these differences
neighbor-joining and maximum parsimony tree con- can probably be attributed to differences in tissue sam-
struction methods clearly divides the SLC6 family into pling between the two species, but many cases are not
four subgroups (Fig. 1). One hundred bootstrap replicas likely to be explained by this. One such example is the
were used in each method. We have chosen to designate absence of the expression of SLC6A1 and SLC6A9 in
these subgroups GABA (bootstrap value: 100, 100, and human eye, while several other transporters such as
52), Monoamine (100, 100, and 94), Amino Acid (98, 97, SLCA8 and SLCA15 have a similar expression in the
and 94) and Orphans (98, 100, and 94) according to sub- eye of the two species. This indicates that there is an
strate preferences. These high bootstrap values and the actual difference in the expression levels that is unlikely
fact that all methods suggest the same topology support to be triggered due to sampling differences.
this subdivision of the SLC6 family into four clades. To obtain an expression profile over the entire family
of SLC6 proteins in rat, we used reverse-transcriptase
EST expression pattern in mouse and human (RT) PCR on mRNA from 14 different rat tissues
(Table 3). The PCR products were analyzed using
The human and mouse EST databases were down- agarose gel electrophoresis and each sample was
loaded and BLAST was performed with the SLC6 pro- analyzed twice where both replicas had to show the
teins as baits. The human and mouse EST sequences same result to be classified as positive. To avoid genomic
were collected and then searched with BLASTX (trans- contamination, all RNA was treated with DNAse and
lated query versus protein database) back against a hu- tested for genomic contamination by excluding RT from
man and mouse RefSeq dataset obtained from http:// the reactions. All cDNAs were tested and verified to
www.ncbi.nlm.nih.gov/. Rat was excluded, as very few work with GAPDH primers and all cDNAs in the panel
ESTs are available from this species. RefSeq is an anno- were diluted to show a similar amount of GAPDH
tated collection of protein sequences and we had added expression. Three of these were CNS tissues, mainly
our novel gene products to this protein dataset to ensure brainstem, cerebellum, and forebrain. Clear expression
that each EST would hit the best corresponding protein. in the brain was seen for 7 of the 20 transporter proteins
This procedure creates a detailed and specific expression but only three Orphan transporters (SLC6A15,
profile for the SLC6 family (Table 2). Five transporters SLCA18, and SLCA19) were found to be abundant in
have a high expression (10 or more ESTs) in the CNS forebrain. Interestingly, the proteins showing high
(mSLC6A1, mSLC6A7, SLC6A8, mSLC6A8, and expression in the forebrain are the two phylogenetically
mSLC6A9) and five transporters have a high expression closely related Orphan transporters SLCA18 and
level in the eye (mSLC6A1, mSLC6A6, SLC6A8, and SLCA19. SLC6A6 is the only one of the proteins that
mSLC6A9, mSLC6A15). Also the Monoamine trans- is expressed in female reproductive organs and our
porters SLC6A2, A3, and A4 are, as expected, mainly EST analysis shows that this is also true for male repro-
expressed in the CNS, although their expression levels ductive organs in mouse.
seem to be relatively low (1, 6, and 2 ESTs for the mouse The expression pattern of three phylogenetically
proteins). The SLC6A2 (16 ESTs) is the only transporter related orphan transporters SLC6A18, SLC6A19, and
with a high expression in female reproductive organs SLC6A20, two of which were shown to be expressed
and this seems to be specific for human, as no ESTs in brain from our RT-PCR panel, was further tested
are detected from this tissue in mouse. mSLC6A6 is for expression in 13 defined sub-dissected regions of
the only transporter that is highly expressed in the male rat brain as well as a number of peripheral tissues using
reproductive organ (20 ESTs) while zero ESTs are real-time quantitative RT-PCR (qPCR). This method
detected for this transcript from testis in human. Three has a higher sensitivity and requires smaller amounts
transporters (mSLC6A8, mSLC6A13, and mSLC6A19) of material than ordinary RT-PCR, making it more suit-
have a high expression in the gastrointestinal system able for sub-dissected brain regions. The qPCR method
and the five (mSLC6A8, mSLC6A13, mSLC6A18, also allows for a quantitative analysis while ordinary
mSLC6A19, and mSLC6A20) are highly expressed in RT-PCR is at best semi-quantitative. In Fig. 2, we dis-
the kidney. These transporters belong to the GABA play the amount of the given transcript relative the
(mSLC6A8 and mSLC6A13) or Orphan (mSLC6A18, amount of GAPDH in 13 brain-regions and eight
mSLC6A19, and mSLC6A20) family. It should be noted peripheral tissues. SLC6A18 and SLC6A19 are ex-
that the GABA family member SLC6A8 has a broad pressed in dorsal, but not ventral, hypothalamus at rel-
expression pattern with high expression in melanocytes atively high levels (50% of GAPDH) while SLC6A20, as
180 P.J. Höglund et al. / Biochemical and Biophysical Research Communications 336 (2005) 175–189

expected from our RT-PCR analysis, does not seem to

be expressed in either of the hypothalamic regions. On
the other hand, certain defined brain regions do have a
substantial expression of this protein, something we
were unable to detect in our larger brain regions. This
is probably due to dilution effects in the less defined
regions.

Evolutionary analysis

We downloaded 13 predicted protein datasets: Taki-

fugu rubripes, Danio rerio, Ciona intestinalis, C. elegans,
Caenorhabditis briggsae, Drosophila melanogaster,
Anopheles gambiae, Arabidopsis thaliana, Schizosaccha-
romyces pombe, and Saccharomyces cerevisiae. They
were searched with the SLC6 HMM and the hits were
retrieved. These proteins were then searched by BLAST
towards an adjusted human Refseq including our novel
human transporters. The proteins were divided into the
GABA, Monoamine, Amino Acid, and Orphan sub-
groups according to the BLAST hits, and all proteins
that did not place into any of these groups were consid-
ered unclassifiable. To be classified as SLC6 proteins,
the five best hits for each protein had to be a SLC6 pro-
tein and four hits had to belong to a specific subgroup in
order to be classified as a member of that subgroup. The
result of the evolutionary analysis is given in Fig. 3. The
numbers of proteins in different groups are similar in
mouse, rat, human, and chicken. The T. rubripes,
D. rerio, and C. intestinalis all had a relatively high
number of GABA transporters. The A. gambiae and
D. melanogaster both had the highest number of trans-
porters in the Amino Acid group whereas most of the
C. elegans proteins are Unclassifiable. The plant (A. tha-
liana) and yeast (S. cerevisiae and S. pombe) predicted
protein datasets do not seem to contain any SLC6-pro-
tein. The predicted protein datasets from the 196 avail-
able bacterial genomes (ftp://ftp.ncbi.nih.gov/genomes/
Bacteria/) yielded 16 GABA, 4 Monoamine, 3 Amino
Acid, 3 Orphans, and 81 Unclassifiable proteins.

Discussion

Our thorough mining of the entire genome of human,

mouse, and rat using HMM led to identification of two
Fig. 2. Expression level of SLC6A18, SLC6A19, and SLC6A20 in
new human SLC6 proteins. We also identified six previ-
different rat tissues relative to GAPDH. Data are presented as ously unidentified rodent proteins. We used these new
means ± SEM (n = 3) as a fraction of GAPDH expression. proteins together with the previously known proteins
Abbreviations: AM, amydala; ACB, acumbens; DHT, dorsal hypo- from this family and performed the first overall evolu-
thalamus; HC, hippocampus; LS, locus ceruleus; PAG, periaqueductal tionary analysis for the entire SLC6-family in mouse,
gray area; PF, prefrontal cortex; PIT, pituitary; SCN, suprachiasmatic
nucleus; SN substantia nigra; VHT, ventral hypothalamus; VP, ventral
rat, and human. We also investigated the number of
pallidum; VTA, ventral tegmental area; ADG, adrenal gland; T, testis; SLC6 sequences in protein prediction datasets from an
H, heart; IT, intestins; K, kidney; L, liver; LU, lung; SM, skeletal additional 13 eukaryotic species. In addition to the phy-
muscle; U, uterus. logenetic and sequence analysis, we also performed an
expression analysis over the entire family of SLC6
 P.J. Höglund et al. / Biochemical and Biophysical Research Communications 336 (2005) 175–189 181

Fig. 3. Evolutionary tree with the number of solute carriers family 6 (SLC6) in different species indicated in graphs. The graph displays the number
of SLC6 genes at the Y-axis and the main class above the X-axis. The number at the nodes indicates the time in million years since the split at that
node occurred, based on fossil data according to [38] +[39] ++[40]. Phylogenetic analyses and subsequent BLAST divides the gene products into five
subgroups: GABA, Monoamine (Monoa.), Amino Acid (AA), Orphans, and Unclassifiable.

proteins, in human and mouse based on EST data and in of their function, structure, and evolution. Our analysis
rat by RT-PCR and qPCR. shows that the SLC6 family in mouse and rat is identi-
Tremendous amount of primary sequence informa- cal. Human and rodents have similar repertoire,
tion from the human, mouse, and rat genomes has been although there are differences in the Orphan group.
released, providing an almost full coverage of the respec- SLC6A16 exists in the human, but not in rodents. The
tive genomes. We can therefore conclude that the gene X transporter protein 3 similar 1(Xtrp3s1) is a duplica-
repertoires presented in Table 1 are fairly complete tion of SLC6A20 that only exists in mouse and rat. Here
and that this is likely to be a complete set of genes from we present two new genes, SLC6A17 and SLC6A18,
the SLC6 family in human, mouse, and rat. The comple- that belong to the Orphan subgroup. SLC6A18 forms
tion of the sequencing projects has shifted the focus a phylogenetic cluster with SLC6A19 and SLCA20
from sequencing and initial protein prediction to the dis- (Fig. 1), while SLC6A17 groups with SLC6A15 and
covery and annotation of new genes and understanding SLC6A16. Our expression studies, based on EST, RT-
182 P.J. Höglund et al. / Biochemical and Biophysical Research Communications 336 (2005) 175–189

Table 1
Description of the complete repertoire of the SLC6 family in human, mouse, and rat
Name Alias Length (aa) Accession No. Chromosomal position [UCSC] Exons
SLC6A1 GAT, GABATR, GABATHG 599 NP_003033 3p25.3 14
SLC6A2 NET, NAT1, NET1 617 NP_001034 16q12.2 13
SLC6A3 DAT, DAT1 620 NP_001035 5p15.33 13
SLC6A4 5-HTT, SERT 630 NP_001036 17q11.2 13
SLC6A5 NET1, GLYT2 797 NP_004202 11p15.1 15
SLC6A6 TAUT 619 NP_003034 3p25.1 13
SLC6A6p 21q21.1
SLC6A7 PROT 636 NP_055043 5q33.1 13
SLC6A8 CT1; CRTR 635 NP_005620 Xq28 13
SLC6A9 GLYT1 692 NP_964012 1p34.1 14
SLC6A10pA 16p11.2
SLC6A10pB 16p11.2
SLC6A11 GAT3; GAT-3 632 NP_055044 3p25.3 14
SLC6A12 BGT1; BGT-1 614 NP_003035 12p13.33 14
SLC6A13 GAT2; GAT-2 602 NP_057699 12p13.33 14
SLC6A14 OBX; ATB(0+) 642 NP_009162 Xq23 14
SLC6A15 V7-3; NTT73; hv7-3 730 NP_877499 12q21.31 11
SLC6A16 NTT5 737 NP_054756 19q13.33 12
SLC6A17 NTT4 727 XP_371280* 1p13.3 11
SLC6A18 Xtrp2; FLJ31236 628 NP_872438 5p15.33 12
SLC6A19 634 NP_001003841 5p15.33 12
SLC6A20 XT3; Xtrp3 592 NP_071800 3p21.31 11
SLC6A21p 19q13.33
mSLC6A1 Gat1; XT-1; Gabt1; Xtrp1 599 P31648 6qE3 14
mSLC6A2 NET 617 NP_033235 8qC5 13
mSLC6A3 DAT, DAT1 619 NP_034150 13qC1 12
mSLC6A4 Htt; Sert; 5-HTT 630 NP_034614 11qB5 13
mSLC6A5 Glyt2 791 NP_683733 7qB3 15
mSLC6A6 Taut 621 NP_033346 6qD1 13
mSLC6A7 Prot 637 NP_958741 18qE1 13
mSLC6A8 CRT; CT1; CRTR; Creat 640 NP_598748 XqA7.1 13
mSLC6A9 Glyt1; Glyt-1 633 NP_032161 4qD2.1 12
mSLC6A11 GAT4 627 B44409 6QE3 14
mSLC6A12 GAT2; Gabt2 614 NP_598422 6qF1 14
mSLC6A13 GAT3; Gabt3 602 NP_653095 6qF1 14
mSLC6A14 ATB0plus; CATB0plus 638 NP_064433 XqA2 14
mSLC6A15 729 NP_780537 10qD1 11
mSLC6A17 NTT4 724 NP_758475 3QF2.3 11
mSLC6A18 XT2; Xtrp2 615 NP_035860 13qC1 12
mSLC6A19 B<0>AT1 634 XP_127449 13qC1 12
mSLC6A20 635 NP_035861 9qF4 11
mXtrp3s1 592 NP_631881 9qF4 11
rSLC6A1 RGD:620533, Gabt1 599 NP_077347 4q42 14
rSLC6A2 RGD:621822, Net 597 NP_112633.1 19p11 14
rSLC6A3 RGD:3715, Dat1 619 NP_036826 1p11 13
rSLC6A4 RGD:3714 653 NP_037166 10q26 15
rSLC6A5 RGD:621824, GLYT2 799 NP_976079 1q22 14
rSLC6A6 RGD:61912 621 NP_058902 4q34 13
rSLC6A7 RGD:620928, Prot 661 NP_446448 18q12.1 13
rSLC6A8 RGD:619711, CHOT1 635 NP_059044 Xq37 13
rSLC6A9 RGD:621243, Glyt1 633 NP_446270 5q36 12
rSLC6A11 RGD:628737, Gabt4, Gat3 627 NP_077348 4q42 14
rSLC6A12 RGD:620255, RNU28927 628 NP_059031 4q42 11
rSLC6A13 RGD:620788, GAT-2 602 NP_598307 4a42 14
rSLC6A14 640 XP_233305 Xq12 14
rSLC6A15 RGD:628664, Ntt73 729 NP_758824 7q21 11
* *
rSLC6A17 727 P31662
rSLC6A18 RGD:69352, Rosit, Xtrp2 615 NP_058859 1p11 12
rSLC6A19 634 1p11 12
rSLC6A20 591 8q32 11
Xtrp3s1 616 Q64093 8q32 11
The columns are name, alias, length in amino acids, accession number, and chromosomal position (from the UCSC website) and the total number of exons.
The ÔrÕ prefix denotes rat and the ÔmÕ prefix indicates mouse. The p suffix as in for instance SLC6A10pA shows that it is a pseudogene. A * means that the
information is not available due to missing genomic information.
 P.J. Höglund et al. / Biochemical and Biophysical Research Communications 336 (2005) 175–189 183

Table 2
Description of the expressed sequence tags (EST) expression in mouse (A) and human (B)
CNS Eye Female reproduction Gastrointestine Kidney Male reproduction Melanocyte Other Stem cells Tumor
A
mSLC6A1 44(30) 12(1) 1
mSLC6A2 1 8(8)
mSLC6A3 6(1) 1
mSLC6A4 2(2) 4 1 2(1)
mSLC6A5 3
mSLC6A6 7(3) 13 2 1 10 20(8) 1 8
mSLC6A7 17(2) 1
mSLC6A8 10(5) 14(10) 7 11 10 19(12) 3 12
mSLC6A9 12(2) 14(2) 4(1) 7(4) 5
mSLC6A11 5(1) 1 1 2(1)
mSLC6A12 1 4 8 1 2(2)
mSLC6A13 7(2) 2 19 15 2(2) 4(2)
mSLC6A14 1(1) 1 8 5(4) 1
mSLC6A15 9(4) 10(7) 4(3) 3
mSLC6A17 8(6) 2 1(1) 1 1
mSLC6A18 3 60 1
mSLCA19 1 11 16 6(1)
mSLC6A20 1 41 2 3(1)
mXtrp3s1 7 1 2 4 1
B
SLC6A1 9 1(1) 1(1) 1(1)
SLC6A2 16 3 12
SLC6A3 1 1
SLC6A4 7
SLC6A5 1 3
SLC6A6 1 1 3 13 3 3
SLC6A7 3
SLC6A8 13(1) 14 8 5 2 7 13 30(2) 6 102
SLC6A9 2 2 4 1 14
SLC6A11 4(1) 1 25 1
SLC6A12 2 3 1 1 4
SLC6A13 3(1) 8 4 5 12 1 1
SLC6A14 1 5 12 7
SLC6A15 2(2) 4 3 2 7 2 6
SLC6A16 7 7 8(1) 1 4
SLC6A17 1 2 4
SLC6A18 1 2 4
SLC6A19
The parenthesis shows the number ESTs from that category found in fetal tissue.

PCR, and qPCR, show that transcripts from the two with the exception of the GABA subgroup in the
novel genes, as well as their phylogenetic neighbors, maximum parsimony analysis where it is 52. It is a
are present in a number of both central and peripheral well-known fact that maximum parsimony generally
tissues, indicating that these genes are likely to partici- performs worse than the other methods for larger data-
pate in several important physiological functions. sets when the degree of sequence identity is low. Interest-
The phylogenetic analysis (Fig. 1) divides the SLC6 ingly, the Amino Acid substrate subgroup is
family into four subgroups which fit remarkably well phylogenetically closer to the Orphan subgroup and
with the known substrate preferences for the SLC6 pro- recently the SLC6A19 gene from the Orphan subgroup
teins, and we have hence opted for naming the phyloge- has been found to transport neutral amino acids
netic branches accordingly. Our phylogenetic analysis [27,28]. Hence, considering that the phylogenetic cluster-
reveals that the human SLC6 repertoire consists of 6 ing for all other groups is according to substrate prefer-
GABA, 3 Monoamine, 4 Amino Acids, and 6 Orphan ence, it is possible that other Orphan SLC6 transporters
proteins, in total of 19 SLC6 proteins. The bootstrap could have amino acids as substrates. As some of these
values that designate the four groups are generally high, Orphan transporters are highly expressed in brain, it is
regardless of the phylogenetic method which shows that tempting to speculate that they function as transporters
the phylogenetic groupings are very stable. The values for amino acids into the cells that are subsequently used
are between 97 and 100 for the four nodes in all cases, as substrates for synthesis of neurotransmitters.
184 P.J. Höglund et al. / Biochemical and Biophysical Research Communications 336 (2005) 175–189

Table 3
RT-PCR analysis of the known SLC6 family on a panel of mRNA from 15 rat tissues
Cerebellum Brainstem Forebrain Uterus Ovary Adrenal Fat Kidney Lung Heart Liver Testis Skeletal Spleen Intestine
gland tissue muscle
rSLC6A1 + +
rSLC6A2 + + + +
rSLC6A3 + +
rSLC6A4 + + + + + +
rSLC6A5 +
rSLC6A6 + + + +
rSLC6A7
rSLC6A8 + + + +
rSLC6A9 + + + +
rSLC6A11
rSLC6A12 +
rSLC6A13 + + +
rSLC6A14 + +
rSLC6A15 + + + + + +
rSLC6A17 + + + + + +
rSLC6A18 + + +
rSLC6A19 + + + +
rSLC6A20 + +
rXtrpP3s1 + +
Dark shaded areas indicate strong expression, whereas + indicates lower expression levels as determined by visual inspection on agarose gels.

Our evolutionary analysis based on GENSCAN pro- ana) and yeast (S. cerevisiae and S. pombe) predicted
tein predictions (Fig. 3) resulted in 22 SLC6 proteins in protein dataset did not contain any SLC6-proteins
human; 8 GABA, 3 Monoamine, 3 Amino Acids, and 7 while bacteria had a substantial number. A total of 16
Orphans, and 1 Unclassified. The total number of GABA, 4 Monoamine, 3 Amino Acid, 3 Orphans,
known human genes is actually 22, but it is divided as; and 81 Unclassifiable bacterial proteins were found in
6 GABA, 3 Monoamine, 3 Amino Acids, and 6 Or- the different genomes. The repertoire of SLC6 proteins
phans. We investigated this discrepancy with the phylo- in C. elegans and insects is rather different than in mam-
genetic analysis and chromosomal alignments. In 12 mals. These do not have any genes belonging to the Or-
cases, there is a 1:1 match between the GENSCAN phan subgroup, which suggests that this subgroup has
predictions and known genes, in three cases two appeared after the divergence of the arthropoda line
GENSCAN genes partially match one known gene, from the lineage leading to vertebrates. Although these
i.e., GENSCAN has erroneously divided one gene into lines do not have an Orphan subgroup the other three
two predictions. In addition, 4 GENSCAN genes groups, GABA, Monoamine, and Amino Acids, are
matched known pseudogenes and four of the known clearly present. Notably, there exist a few Unclassifiable
genes were not present in the gene prediction dataset. genes in the genome of most species. For example there
There is on average about 14 introns per human SLC6 are 6 Unclassifiable genes in A. gambiae and 13 in
protein, and this complex exon–intron arrangement C. elegans and 81 in bacteria. This is very intriguing
makes it hard for ab initio programs such as GEN- and it is likely that one or more phylogenetic subgroups
SCAN to make correct predictions. This kind of analysis either appeared specifically in these lineages or existed
of GENSCAN genes, especially in gene families with a in parallel with the three subgroups common to verte-
high number of introns, can be used to estimate the pro- brates and invertebrates, but were lost before the
portions of proteins in each subgroup but not to predict appearance of vertebrates.
the absolute number of genes. In rat, nine of the 19 SLC6 proteins have high expres-
Our analysis shows that SLC6 genes are present in all sion in brain, while three have low expression there. Of
animal species investigated, but not in plants or fungi. the known monoamine transporters for dopamine
The repertoire of SLC6A proteins, as estimated from (SLC6A3), serotonin (SLC6A4), and noradrenaline
GENSCAN proteins, for mouse, human, rat, and chick- (SLC6A2) only the serotonin transporter can be detect-
en are very similar. However the T. rubripes, D. rerio, ed in large brain regions while all three are likely be
and C. intestinalis all had a 2- to 3-fold higher number detected in more defined regions. The three phylogenet-
of transporters in the GABA group. The A. gambiae ically related GABA transporters SLC6A1, SLC6A8
and D. melanogaster both had most transporters in and SLC6A6 were found in the CNS, while the other
the Amino Acid group while C. elegans had the highest three proteins from the GABA group, SLC6A11,
number in the unclassifiable group. The plant (A. thali- SLC6A12, and SLC6A13, seem to be less highly ex-
 P.J. Höglund et al. / Biochemical and Biophysical Research Communications 336 (2005) 175–189 185

pressed and have more expression in the periphery. they code for functional proteins. Moreover, we provide
Expression in the periphery for rat GABA transporters expression profiles for the entire repertoire SLC6
has previously been shown for SLCA12 [32] and mRNA in rats which may give better insight into the
SLC6A13 [33]. In a similar manner, we do notice a high possible physiological role of these genes.
expression of all three monoamine transporters in
peripheral tissues in rat, which is also supported by
the EST data from human and mouse. This could be Materials and methods
one reason for the side effects caused by antidepressant
drugs in these tissues. The Amino Acid transporters Collecting the original dataset
SLC6A5 and SLC6A9 are expressed in brain and
Seventeen known human SLC6 sequences were initially identified
peripheral tissue, while SLC6A7 and SLC6A14 are ex-
on the HUGO Gene Nomenclature Committee Homepage (http://
pressed only in the periphery. Recently, polymorphisms www.gene.ucl.ac.uk/nomenclature/). Fourteen mouse sequences were
in the SLC6A14 gene were shown to be involved in initially identified on the Mouse Genome Informatics web page (http://
inherited forms of obesity [11,12].The gene is expressed www.informatics.jax.org/). The rat genome database RatMap (http://
in the gastrointestinal region in human and mouse with ratmap.gen.gu.se) only contained six SLC6 sequences and therefore we
also searched Rat Genome Database (http://rgd.mcw.edu/SSS) which
EST analysis and in fat tissue in rat. The transporters
contained eight more genes which resulted in a total of 14 unique rat
from the Orphan group are so called because the major- genes. All the rat, mouse, and human protein sequences were thereafter
ity of them do not have any known substrate. Recently, downloaded from the Entrez web page (http://www.ncbi.nlm.nih.gov/
human and mouse SLC6A19 was shown to be able to entrez/). We also searched UniGene on the NCBI website and the
transport neutral amino acids [27] and mutations in this literature for known genes.
gene have been shown to be the cause of the renal dis-
Identification of novel genes
ease Hartnup disorder. This disease is an autosomal
recessive disorder which results from impaired transport The collected SLC6 proteins were used as baits for searches in both
of neutral amino acids across epithelial cells in renal the human, mouse, and rat protein databases. Information was col-
proximal tubules and intestinal mucosa. It has also been lected from both NCBI (http://www.ncbi.nih.gov/) and the Celera
proposed that the allelic heterogeneity of the Orphan Genomics (http://www.celeradiscoverysystem.com/) databases. The
N- and C-termini were removed after being identified through searches
transporter SLC6A16 may explain the variation in the
against the RPS-BLAST database (http://www.ncbi.nlm.nih.gov/
biochemical and clinical phenotype in pedigrees with Structure/cdd/wrpsb.cgi). The truncated genes were aligned using
Hartnup disorder [28]. The SLC6A19 in rat shows ClustalW 1.83 with default parameters applied [34]. From the align-
expression in several peripheral tissues including intes- ment, a HMM model was constructed with hmmbuild from the 2.3.2
tine and in mouse this gene seems to have high expres- release of the HMMER package [29]. The model was calibrated using
hmmcalibrate with default parameters. Searching was conducted in
sion in intestine and kidney. On the other hand, our
three human, one rat, and two mouse databases with hmmsearch using
human EST panels indicate that this gene has a very E = 1e 4 as the cut-off value. For human we searched the Gnomon
low expression level as not a single EST is found in protein data set (Build 34 version 3) and the GENSCAN protein
any tissue. Also, we do not find any evidence of expres- dataset (assembly 28), both downloaded from NCBI ftp site (ftp://
sion of the SLC6A16 protein in colon or kidney. The ftp.ncbi.nih.gov/genomes/). In addition, we also searched the GEN-
SCAN dataset from Ensembl based on the NCBI version 34 assembly.
fact that the gene for SLC6A16 is absent in both mouse
For mouse, we used the Gnomon protein dataset of the NCBI Mouse
and rat complicates further the use of rodents as a model build 32 as well as the ab initio Ensembl (http://www.ensembl.org/)
for this disease. The Orphan transporter SLC6A18 is version based on the NCBI Mouse build 32. For rat, we searched the
highly expressed in brain and we can also detect Gnomon protein dataset, which is based on the NCBI build 2 from the
SLC6A20 in certain subdissected brain regions of rat. Rat Genome Sequencing consortium (RGSC) v3.1 assembly.
We used protein–protein (BLASTP) searches of the non-redundant
Also SLC6A15 and SLC6A17 as well as the rodent spe-
protein database at NCBI (http://www.ncbi.nih.gov/BLAST/), with
cific Xtrp3s1, are expressed in the brain of all species the previously downloaded SLC6-genes as baits [35]. BLAST searches
investigated. This is intriguing as some of these proteins against human, mouse, and rat proteins were performed, without fil-
could have important functions in the brain such as act- tering for low complexity. The best five alignments for each search
ing as transporters for amino acids used in the synthesis were manually inspected.
All the sequences were searched using BLAT against the human,
of neurotransmitters or being specific reuptake proteins
mouse, and rat database at UCSC Genome Bioinformatics Site (http://
for neurotransmitters which currently do not have a genome.ucsc.edu). This was carried to both a species-to-species man-
known such system, for example trace amines. ner as well as across species. All high-scoring hits were visually
In this analysis, we have provided a comprehensive inspected in order to find previously unidentified family members of
overview of the repertoire of SLC6 genes in a number the SLC6 family.
of species. This provides the molecular foundation for
Verification of novel genes
one of the important groups of targets for drugs that
act on the transport uptake and flow of various neuro- The novel transporters were searched with BLASTP and RPS-
transmitters, amino acids, and osmolytes. We have iden- BLAST. All genes with a SLC6 gene as the best BLASTP hit were
tified new genes and provide convincing evidence that subsequently searched with RPS-BLAST. The search with RPS-
186 P.J. Höglund et al. / Biochemical and Biophysical Research Communications 336 (2005) 175–189

BLAST had to yield a PFAM00209 (sodium:neurotransmitter sym- after searched with BLASTX against a human and mouse RefSeq
porter family) model as the best hit. All proteins that did not match (ftp://ftp.ncbi.nih.gov/refseq/) datasets, respectively. We had added
both these two criteria were excluded. our novel transporters to these protein datasets to ensure that each
EST would hit the best corresponding protein. We removed all ESTs
Identification of EST-clones that did not have a SLC6-gene as first hit. We then used in-house
scripts to extract relevant expression information from the EST
The GENSCAN software, used for predicting coding regions for database. The ESTs were sorted and entered into a spreadsheet for
manual inspection.
the human genome project, has the capacity to predict approximately
80% of the splice sites correctly [36]. Therefore, the coding regions
needed to be verified. We used mRNA and expressed sequence tags Description and retrieval of predicted protein datasets
(ESTs) from mouse, human, and rat, respectively. BLAT searches at the
UCSC web site were used to obtain genomic nucleotide sequences for The human, mouse, rat, D. rerio, G. gallus, A. gambiae, T. rubripes,
the novel protein sequences. The full genomic DNA sequences of the D. melanogaster, and C. elegans predicted protein datasets were all
novel GPCRs were searched against the human and mouse NCBI EST retrieved from the Ensembl website. Ensembl predicted peptides are
databases, respectively. There is no rat-specific database and therefore available in two different sets. One is the ‘‘pep’’ version, which consists
we searched the entire EST database. The search was limited by the of the super-set of all translations from known, ab initio or novel gene
Entrez query ‘‘Rattus norvegicus [ORGN]’’. BLASTN against the predictions. This set uses biological information such as EST-support
nucleotide non-redundant database with the limit biomol_mrna[prop- to produce several different transcripts from each gene. The other set is
erties] was used to find the corresponding mRNA sequences. The the ab initio dataset with all predictions based entirely on the genomic
alignments with the identified (ESTs) and cDNA sequences were sequences and no biological evidence. Prediction algorithms, such as
manually inspected to ensure correct identity and exclude non-relevant SNAP and GENSCAN, have been used to create the datasets. The
hits. The predicted coding regions were verified by assembling the EST ab initio approach does not distinguish between pseudogenes and
sequences and the full genomic DNA sequences using SeqMan 5.01 biologically real proteins. We however decided to use ab initio pre-
from the DNASTAR package. The genomic DNA sequences were dicted protein datsets. Ab initio is a more honest and unbiased ap-
considered correct, and the EST and mRNA sequences were only used proach, because the result of the analyzes is not influenced by the
to correct the predicted exon–intron boundaries. EditSeq 5.02 from amount of ESTs or any other biological data available. Thus, the
DNASTAR package was used to find the right open reading frames and ab initio set is better to use in a cross-species comparison. Ab initio
translate the nucleotide sequences into final novel protein sequences. also seems like a superior dataset in our test runs. The D. melanogaster
dataset did only have a ‘‘pep’’ version. We used this version, but re-
moved the multiple translated protein version and left only one tran-
Phylogenetic analysis
script for each gene. The sources for the datasets for the yeast and
bacterial genomes as well as the datasets for A. thaliana and
Our protein sequences were all aligned with ClustalW version 1.83
C. intestinalis are indicated below. All other datasets were downloaded
[34]. The alignments were bootstrapped 100 times with SEQBOOT
from the Ensemble site. Below follows a description of these datasets.
from the Linux version of the PHYLIP 3.6 package [37]. Three dif-
Anopheles gambiae. The ensemble A. gambiae (mosquito) genome
ferent methods, neighbor-joining, maximum parsimony, and maxi-
release 17.2a.1 is a re-annotation of the second version of the
mum likelihood, were used to make phylogenetic trees from the
A. gambiae genome assembly.
SEQBOOT outfile. For the neighbor-joining method, protein distances
Arabidopsis thaliana. The Arabidopsis predicted protein dataset was
were calculated with PROTDIST using the Jones–Taylor–Thornton
downloaded from ftp://ftp.arabidopsis.org/home/tair/Sequences/
matrix. The trees were calculated on the 100 different distance matrixes
blast_datasets/.
previously generated with PROTDIST, using NEIGHBOR.
Bacteria. The predicted protein datasets from the 196 available
CONSENSE from the Linux version of the PHYLIP 3.6 package was
bacterial genomes were downloaded from ftp://ftp.ncbi.nih.gov/ge-
used to obtain a bootstrapped consensus tree.
nomes/Bacteria/. The genomes have been searched with three predic-
Maximum parsimony trees were calculated from the previously
tion methods; GeneMark, GeneMark.hmm, and Glimmer. Fourteen
derived SEQBOOT file using PROTPARS from the Linux version of
genomes have also been reviewed by NCBI experts: Aeropyrum pernix,
the PHYLIP 3.6 package. The trees were unrooted and calculated
Archaeoglobus fulgidus, Buchnera sp, Buchnera aphidicola, Corynebac-
using ordinary parsimony and the topologies were obtained using the
terium glutamicum, Haemophilus influenzae, Mycoplasma genitalium,
built-in tree search procedure. CONSENSE was used to obtain a
Mycoplasma pneumoniae, Oceanobacillus iheyensis, Shewanella oneid-
bootstrapped consensus tree. Maximum likelihood trees were calcu-
ensis, Pyrococcus abyssi, Pyrococcus horikoshii, Thermoplasma volca-
lated, from the previously derived SEQBOOT file, using PROML from
nium, and Vibrio vulnificus.
the Linux version of the PHYLIP 3.6 package. PROML was used with
Caenorhabditis elegans. The v25.116a.1 ensembl predicted protein
default settings and CONSENSE was used to obtain a bootstrapped
dataset of C. elegans is a direct import of the Wormbase (http://
consensus tree. All trees were plotted using TreeView (http://taxonomy.
www.wormbase.org/) 116 dataset. No additional gene building pro-
zoology.gla.ac.uk/rod/treeview.html).
cedures have been preformed.
Gallus gallus. The chicken (G. gallus) retrieved is the ensembl
Expression profile 25.1b.1 version, based on the Chicken 1a build.
Ciona intestinales. The ciona proteins were downloaded from
The EST database at NCBI currently has a total of 24.6 million http://genome.jgi-psf.org.
entries. 6.0 million, 4.3 million, and 0.7 million are human, mouse, and Danio rerio. The genome of zebrafish (D. rerio) used was the
rat EST sequences, respectively (http://www.ncbi.nlm.nih.gov/dbEST/ ensembl zebrafish release 24.4.1 based on the zebrafish assembly ver-
index.html). The rat EST database was considered too small for sion 4 (Zv4) produced by the Wellcome Trust Sanger Institute (http://
making a rat EST expression profile. We searched the whole human www.sanger.ac.uk/Projects/D_rerio/).
and mouse EST databases with our SLC6 protein sequences as input Drosophila melanogaster. The ensembl 25.3b.1 version of D. mela-
files. We used protein query versus translated database, TBLASTN, nogaster genome is based on the the Flybase (http://www.flybase.org/)
from the BLAST 2.2.9 package. A cut-off value of 0.1 was used. The 3.1 assembly. It is not available as an ab initio dataset.
human and mouse EST sequences were collected and then searched Homo sapiens. We used the human ab initio ensembl v21.34d.1
with BLASTX (translated query versus protein database) and there- build which is based on the NCBI 34 assembly of the human genome.
 P.J. Höglund et al. / Biochemical and Biophysical Research Communications 336 (2005) 175–189 187

Mus musculus. We used the mouse ab initio release 24.33.1 based Reverse transcriptase PCR
on the NCBI m33 mouse assembly.
Rattus norvegicus. We used the ensembl ab initio rat v25.3c.1 The presence of individual transcripts in certain tissues was deter-
assembly. This version is based on the draft genome assembly (v3.1) mined using reverse transcriptase (RT) PCR. RT-PCR was performed
found on the Baylor College of Medicine Human Genome Sequencing in a 25 ll final reaction volume using a Perkin Elmer 9700 thermal
Center (http://www.hgsc.bcm.tmc.edu/projects/rat/). cycler (Applied Biosystems, Stockholm, Sweden). The PCR included
Saccharomyces cerevisiae. The budding yeast (S. cerevisiae) has a 20 mM Tris–HCl (pH 8.4), 50 mM KCl, 4 mM MgCl2, and 0.2 mM
current assembly available from the Sanger institute ftp-site (ftp:// dNTP. Template concentration was 4 ng/ll, and the concentration of
ftp.sanger.ac.uk/pub/yeast/cerevisiae/cerpep/). The predicted protein each primer was 1 pmol/ll. Taq DNA polymerase (Invitrogen) was
dataset consists of 5803 proteins. used at 0.04 U/ll. Annealing temperature was 55 °C and 35 cycles were
Schizosaccharomyces pombe. The fission yeast (S. pombe) is also performed. All experiments were repeated twice. The PCR products
available from Sanger institute ftp-site(ftp://ftp.sanger.ac.uk/pub/ were analyzed using agarose gel electrophoresis and scored manually.
yeast/pombe/Protein_data/). The GENSCAN dataset consists of 4983 A certain tissue/primer pair combination had to be positive on both
proteins. occasions to be considered positive. All PCR-primers were designed
Takifugu rubripes. The ensembl fugu (T. rubripes) genome release using the Primer 3 (http://frodo.wi.mit.edu/primer3/primer3_code.
v25.2c.1 is based on the second Fugu assembly from the Fugu Con- html) software using default settings, with a required product length
sortium (http://www.fugu-sg.org/project/info.html). between 80 and 110 nucleotides. All primer sequences are available
from the authors upon request.
Identification of SLC6 sequences in the evolutionary analysis
Real-time quantitative reverse transcriptase PCR
The predicted protein datasets previously retrieved were used.
These predicted protein datasets were searched against a HMM-model Relative levels of mRNA were determined by quantitative reverse
(pfam00209.11) downloaded from http://www.sanger.ac.uk/cgi-bin/ transcriptase PCR (qPCR). qPCR was performed in a 25 ll final
Pfam/getacc?PF00209 and searched using HMMsearch from the 2.3.2 reaction volume using an iCycler real-time detection instrument (Bio-
release of the HMMER package using E = 10 as cut-off value. The hits Rad Laboratories, Sundbyberg, Sweden). The reaction contained
in the result file were used to extract Fasta files, with an in-house script 20 mM Tris–HCl (pH 8.4), 50 mM KCl, 4 mM MgCl2, 0.2 mM dNTP,
based on fastacmd from the NCBI BLAST-suite, from the original and SYBR Green (1:50,000). Template concentration was 1 ng/ll, and
protein prediction file. the concentration of each primer was 0.8 pmol/ll. Taq DNA poly-
merase (Invitrogen) was used at 0.02 U/ll. All samples were measured
Subdivision of the SLC6 proteins in triplicate and compared with a no-template control for each primer
pair. Annealing temperature was 62 °C and 50 cycles were performed.
We divided our collection of mouse, human, and rat SLC6 proteins Melting point curves were included to confirm that only one product
into five groups: (I) GABA & others (GABA), (II) Monoamine, (III) was formed. For a product to be classified as positive the CT-value for
Amino acid, (IV) Orphan, and (V) Unclassifiable. This was done that product had to be two units higher than the corresponding neg-
according to the phylogenetic analyses previously described. A file with ative control and contain only one peak in the melting curve.
all human RefSeq genes was downloaded from the ncbi ftp site: ftp://
ftp.ncbi.nih.gov/refseq/H_sapiens/mRNA_Prot/. The RefSeq (Refer-
ence Sequence) protein is an annotated collection of non-redundant
proteins. We thereafter added our confirmed 19 mouse and 19 rat
Acknowledgments
proteins, and adjusted the RefSeq to make the human SLC6 repertoire
look exactly as in our own dataset. We thereafter searched all our We thank Dr. Helgi B Schiöth (supported by the
predicted proteins from the HMM search with BLASTP with an e- Swedish Research Council) for valuable discussions
value of e 12. To be classified as SLC6 proteins, the five best hits for regarding the manuscript. The studies were supported
each protein had to be a SLC6 protein. The first four hits had to be-
long to a specific subgroup in order to be classified as a member of the
by the Swedish Society for Medical Research (SSMF),
subgroup. If, for instance, the first three BLAST hits were GABA Svenska Läkaresällskapet, Åke Wikberg Foundation,
proteins and hits four and five were Monoamine, the protein was Thurings Foundation, and Magnus Bergwalls
concluded to be a Unclassifiable SLC6 protein. Foundation.

RNA isolation and cDNA synthesis

Individual tissue samples were homogenized by sonication in Appendix A. Supplementary data

TRIzol reagent (Invitrogen) using a Branson sonifier. Chloroform was
added to the homogenate, which was then centrifuged at 12,000 rpm Supplementary data associated with this article can
for 15 min at 4 °C. The water phase was transferred to a new tube, and
RNA was precipitated with isopropanol. The pellets were cleaned with
be found, in the online version, at doi:10.1016/
75% ethanol, air-dried at room temperature, and dissolved in RNAse- j.bbrc.2005.08.048.
free water. DNA contamination was removed by treatment with
DNAse (Roche Diagnostics) for 12 h at 16 °C. The DNAse was then
inactivated by heating the samples to 75 °C for 15 min. RNA purity
was checked by PCR, using a primer for GAPDH (see below), and References
RNA concentration was determined with an Ultrospec 2100 spectro-
photometer (Amersham Biosciences). cDNA was synthesized with re- [1] J.C. Venter, M.D. Adams, E.W. Myers, P.W. Li, R.J. Mural,
verse transcriptase from Amersham Biotech, using random hexamers G.G. Sutton, H.O. Smith, M. Yandell, C.A. Evans, R.A. Holt,
as primers and following the instructions of the manufacturer. The et al., The sequence of the human genome, Science 291 (2001)
quality of the cDNA was confirmed with PCR as above. 1304–1351.
188 P.J. Höglund et al. / Biochemical and Biophysical Research Communications 336 (2005) 175–189

[2] R.H. Waterston, K. Lindblad-Toh, E. Birney, J. Rogers, J.F. some band 16p11.2 by in situ hybridization, Cytogenet. Cell
Abril, P. Agarwal, R. Agarwala, R. Ainscough, M. Alexanders- Genet. 76 (1997) 19.
son, P. An, et al., Initial sequencing and comparative analysis of [21] E.E. Eichler, F. Lu, Y. Shen, R. Antonacci, V. Jurecic, N.A.
the mouse genome, Nature 420 (2002) 520–562. Doggett, R.K. Moyzis, A. Baldini, R.A. Gibbs, D.L. Nelson,
[3] R. Fredriksson, M.C. Lagerstrom, L.G. Lundin, H.B. Schioth, Duplication of a gene-rich cluster between 16p11.1 and Xq28: a
The G-protein-coupled receptors in the human genome form five novel pericentromeric-directed mechanism for paralogous genome
main families. Phylogenetic analysis, paralogon groups, and evolution, Hum. Mol. Genet. 5 (1996) 899–912.
fingerprints, Mol. Pharmacol. 63 (2003) 1256–1272. [22] H. Lill, N. Nelson, Homologies and family relationships among
[4] R. Fredriksson, H.B. Schioth, The repertoire of G-protein Na+/Cl neurotransmitter transporters, Methods Enzymol. 296
coupled receptors in fully sequenced genomes, Mol. Pharmacol. (1998) 425–436.
(2005). [23] T.P. Larsson, C.G. Murray, T. Hill, R. Fredriksson, H.B. Schioth,
[5] G. Manning, D.B. Whyte, R. Martinez, T. Hunter, S. Sudarsa- Comparison of the current RefSeq, Ensembl and EST databases
nam, The protein kinase complement of the human genome, for counting genes and gene discovery, FEBS Lett. 579 (2005)
Science 298 (2002) 1912–1934. 690–698.
[6] S. Caenepeel, G. Charydczak, S. Sudarsanam, T. Hunter, G. [24] R. Fredriksson, P.J. Hoglund, D.E. Gloriam, M.C. Lagerstrom,
Manning, The mouse kinome: discovery and comparative genom- H.B. Schioth, Seven evolutionarily conserved human rhodopsin G
ics of all mouse protein kinases, Proc. Natl. Acad. Sci. USA 101 protein-coupled receptors lacking close relatives, FEBS Lett. 554
(2004) 11707–11712. (2003) 381–388.
[7] F.H. Yu, W.A. Catterall, The VGL-chanome: a protein super- [25] R. Fredriksson, M.C. Lagerstrom, P.J. Hoglund, H.B. Schioth,
family specialized for electrical signaling and ionic homeostasis, Novel human G protein-coupled receptors with long N-terminals
Sci STKE (2004) re15. containing GPS domains and Ser/Thr-rich regions, FEBS Lett.
[8] N.H. Chen, M.E. Reith, M.W. Quick, Synaptic uptake and 531 (2002) 407–414.
beyond: the sodium- and chloride-dependent neurotransmitter [26] R. Fredriksson, D.E. Gloriam, P.J. Hoglund, M.C. Lagerstrom,
transporter family SLC6, Pflugers Arch. 447 (2004) 519–531. H.B. Schioth, There exist at least 30 human G-protein-coupled
[9] M.A. Hediger, M.F. Romero, J.B. Peng, A. Rolfs, H. Taka- receptors with long Ser/Thr-rich N-termini, Biochem. Biophys.
naga, E.A. Bruford, The ABCs of solute carriers: physiological, Res. Commun. 301 (2003) 725–734.
pathological and therapeutic implications of human membrane [27] R. Kleta, E. Romeo, Z. Ristic, T. Ohura, C. Stuart, M. Arcos-
transport proteins: introduction, Pflugers Arch. 447 (2004) Burgos, M.H. Dave, C.A. Wagner, S.R. Camargo, S. Inoue,
465–468. et al., Mutations in SLC6A19, encoding B0AT1, cause Hartnup
[10] E.H. Rosenberg, L.S. Almeida, T. Kleefstra, R.S. deGrauw, H.G. disorder, Nat. Genet. 36 (2004) 999–1002.
Yntema, N. Bahi, C. Moraine, H.H. Ropers, J.P. Fryns, T.J. [28] H.F. Seow, S. Broer, A. Broer, C.G. Bailey, S.J. Potter, J.A.
deGrauw, et al., High prevalence of SLC6A8 deficiency in X- Cavanaugh, J.E. Rasko, Hartnup disorder is caused by mutations
linked mental retardation, Am. J. Hum. Genet. 75 (2004) 97–105. in the gene encoding the neutral amino acid transporter SLC6A19,
[11] E. Suviolahti, L.J. Oksanen, M. Ohman, R.M. Cantor, M. Nat. Genet. 36 (2004) 1003–1007.
Ridderstrale, T. Tuomi, J. Kaprio, A. Rissanen, P. Mustajoki, P. [29] S.R. Eddy, Profile hidden Markov models, Bioinformatics 14
Jousilahti, The SLC6A14 gene shows evidence of association with (1998) 755–763.
obesity, J. Clin. Invest. 112 (2003) 1762–1772. [30] T. Ota, Y. Suzuki, T. Nishikawa, T. Otsuki, T. Sugiyama, R. Irie,
[12] E. Durand, P. Boutin, D. Meyre, M.A. Charles, K. Clement, C. A. Wakamatsu, K. Hayashi, H. Sato, K. Nagai, Complete
Dina, P. Froguel, Polymorphisms in the amino acid transporter sequencing and characterization of 21,243 full-length human
solute carrier family 6 (neurotransmitter transporter) member 14 cDNAs, Nat. Genet. 36 (2004) 40–45.
gene contribute to polygenic obesity in French Caucasians, [31] A. Broer, K. Klingel, S. Kowalczuk, J.E. Rasko, J. Cavanaugh, S.
Diabetes 53 (2004) 2483–2486. Broer, Molecular cloning of mouse amino acid transport system
[13] R.D. Blakely, L.J. De Felice, H.C. Hartzell, Molecular physiology B0, a neutral amino acid transporter related to Hartnup disorder,
of norepinephrine and serotonin transporters, J. Exp. Biol. 196 J. Biol. Chem. 279 (2004) 24467–24476.
(1994) 263–281. [32] C.E. Burnham, B. Buerk, C. Schmidt, J.C. Bucuvalas, A liver-
[14] H. Nelson, S. Mandiyan, N. Nelson, Cloning of the human brain specific isoform of the betaine/GABA transporter in the rat:
GABA transporter, FEBS Lett. 269 (1990) 181–184. cDNA sequence and organ distribution, Biochim. Biophys. Acta
[15] T. Pacholczyk, R.D. Blakely, S.G. Amara, Expression cloning of a 1284 (1996) 4–8.
cocaine- and antidepressant-sensitive human noradrenaline trans- [33] L.A. Borden, K.E. Smith, E.L. Gustafson, T.A. Branchek, R.L.
porter, Nature 350 (1991) 350–354. Weinshank, Cloning and expression of a betaine/GABA trans-
[16] B. Giros, S. el Mestikawy, N. Godinot, K. Zheng, H. Han, T. porter from human brain, J. Neurochem. 64 (1995) 977–984.
Yang-Feng, M.G. Caron, Cloning, pharmacological characteriza- [34] J.D. Thompson, D.G. Higgins, T.J. Gibson, CLUSTAL W:
tion and chromosome assignment of the human dopamine improving the sensitivity of progressive multiple sequence align-
transporter, Mol. Pharmacol. 42 (1992) 383–390. ment through sequence weighting, position-specific gap penalties
[17] G.R. Uhl, S. Kitayama, P. Gregor, E. Nanthakumar, A. Persico, and weight matrix choice, Nucleic Acids Res. 22 (1994) 4673–
S. Shimada, Neurotransmitter transporter family cDNAs in a rat 4680.
midbrain library: Õorphan transportersÕ suggest sizable structural [35] S.F. Altschul, T.L. Madden, A.A. Schaffer, J. Zhang, Z. Zhang,
variations, Brain Res. Mol. Brain Res. 16 (1992) 353–359. W. Miller, D.J. Lipman, Gapped BLAST and PSI-BLAST: a new
[18] J. Masson, C. Sagne, M. Hamon, S. El Mestikawy, Neurotrans- generation of protein database search programs, Nucleic Acids
mitter transporters in the central nervous system, Pharmacol. Rev. Res. 25 (1997) 3389–3402.
51 (1999) 439–464. [36] C. Burge, S. Karlin, Prediction of complete gene structures in
[19] G.S. Iyer, R. Krahe, L.A. Goodwin, N.A. Doggett, M.J. Siciliano, human genomic DNA, J. Mol. Biol. 268 (1997) 78–94.
V.L. Funanage, R. Proujansky, Identification of a testis-expressed [37] J. Felsenstein, PHYLIP Phylogenetic inference package, Distrib-
creatine transporter gene at 16p11.2 and confirmation of the X- uted by the author, Department of genetics, University of
linked locus to Xq28, Genomics 34 (1996) 143–146. Washington, Seattle, WA (1993).
[20] W. Xu, L. Liu, P.A. Gorman, D. Sheer, P.C. Emson, Assignment [38] S. Blair Hedges, S. Kumar, Genomic clocks and evolutionary
of the human creatine transporter type 2 (SLC6A10) to chromo- timescales, Trends Genet. 19 (2003) 200–206.
 P.J. Höglund et al. / Biochemical and Biophysical Research Communications 336 (2005) 175–189 189

[39] P. Dehal, Y. Satou, R.K. Campbell, J. Chapman, B. Degnan, A. De [41] M.K. Farmer, M.J. Robbins, A.D. Medhurst, D.A. Campbell, K.
Tomaso, B. Davidson, A. Di Gregorio, M. Gelpke, D.M. Good- Ellington, M. Duckworth, A.M. Brown, D.N. Middlemiss, G.W.
stein, et al., The draft genome of Ciona intestinalis: insights into Price, M.N. Pangalos, Cloning and characterization of human
chordate and vertebrate origins, Science 298 (2002) 2157–2167. NTT5 and v7-3: two orphan transporters of the Na+/Cl -depen-
[40] M.S. Springer, W.J. Murphy, E. Eizirik, S.J. OÕBrien, Placental dent neurotransmitter transporter gene family, Genomics 70 (2000)
mammal diversification and the Cretaceous-Tertiary boundary, 241–252.
Proc. Natl. Acad. Sci. USA 100 (2003) 1056–1061.