The low molecular mass polypeptide (LMP2, LMP7, and MECL-1) genes code for β-type subunits of the proteasome, a multimeric complex that degrades proteins into peptides as part of the MHC class I-mediated Ag-presenting pathway. These gene products are up-regulated in response to infection by IFN-γ and replace the corresponding constitutively expressed subunits (X, Y, and Z) during the immune response. In humans, the LMP2 and LMP7 genes both reside within the class II region of the MHC (6p21.3), while MECL-1 is located at 16q22.1. In the present study, we have identified all three IFN-γ-regulated β-type proteasome subunits in Fugu, which are present as a cluster within the Fugu MHC class I region. We show that in this species, LMP7, LMP2, and MECL-1 are linked. Also within this cluster is an LMP2-like subunit (which seems specific to all teleosts tested to date) and a closely linked LMP7 pseudogene, indicating that within Fugu and potentially other teleosts, there has been an additional regional duplication involving these genes.

The MHC class I Ag presentation pathway is reliant upon several stages; Ags are proteolytically degraded in the cytosol into peptides 8–9 aa long by proteasomes; these peptide fragments are then transported across the rough endoplasmic reticulum membrane by specialized peptide transporter proteins and loaded into the peptide-binding domain of the class I molecule. The class I/Ag complex is then transported to the cell surface for destruction of the Ag by cytotoxic T cells. Thus, the proteasome plays a pivotal role in the adaptive immune system (1).

The 20S proteasome, which functions as the proteolytic core of the 26S protease complex, has a cylindrical structure with four layers of rings, each composed of seven subunits. The two outer rings are made up of α-type subunits, while the two inner rings are composed of β-type subunits (2, 3). The catalytic sites of the 20S proteasome reside in the β-type subunits. These β-type subunits include Y (also known as δ, 2, low molecular mass polypeptide 19 (LMP19),3 PSMB6), X (ε, MB1, 10, LMP17, PSMB5), Z (86, LMP9, MC14, PSMB7), LMP2 (PSMB9), LMP7 (PSMB8), and MECL-1 (PSMB10, LMP10) (4, 5, 6, 7, 8). In response to infection, IFN γ is produced, which up-regulates the production of LMP2, LMP7, and MECL-1, which displace the constitutively expressed proteasome β-type subunits (9, 10). These changes in subunit composition appear to facilitate class I-mediated Ag presentation by altering the cleavage specificities of the proteasome (11, 12).

Of the β-type subunits identified to date in mammals, the IFN-γ-regulated subunits are most closely related to the constitutive subunits that they replace, i.e., LMP7/X, LMP2/Y, and MECL-1/Z, both in terms of amino acid sequence and genomic structure. Therefore, each exchangeable proteasome subunit pair appears to have emerged by gene duplication from its respective common ancestor. However, the three pairs of β-type subunits do not appear to share an immediate common ancestor. The proteasome β-type subunits are molecules of very ancient origin that are found even in bacteria. By the time the eukaryotes emerged, the ancestral β-type subunit gene had already undergone several successive duplications to produce the X-, Y-, and Z-like subunits.

It is hypothesized that during the evolution of vertebrates, there were two whole genome duplication events, which have resulted in the formation of paralogous regions in the mammalian genome. With particular reference to the β-type subunits, one theory proposes that the original X-, Y-, and Z-like β-type subunits were present in a tandem array. The first duplication event then resulted in the formation of separate X, Y, Z, and LMP2, LMP7, and MECL-1 clusters, which went on to form 6p21.3 and 9q24 in humans (13, 14, 15). This event has been implicated in the formation of the primordial MHC region in vertebrates, with, in humans, the LMP2 and LMP7 genes staying within the MHC and MECL-1 being translocated to 16q22.1. Subunit Z remained in the primordial 9q34 region, with X being translocated to 14q11.2 and Y to 17p13 (13).

The locations of these proteasome subunits are syntenic in human and mouse (7), which are the only species with detailed mapping data available. Only partial data exist to date in other fish species; linkage of LMP2 and LMP7 has been demonstrated in medaka (16) and the LMP2/TAP locus has been described in rainbow trout (Onchorhynchus mykiss, Omy) (17). Recently, a 26-kb proteasome subunit cluster was described in another teleost, zebrafish (Danio rerio, Dre) comprising LMP2, PSMB11 (belonging to the Y proteasome evolutionary-derived group), PSMB12 (belonging to the Z proteasome-derived group), LMP7, and TAP2 (18). This finding is exciting and will be discussed for its evolutionary implication.

In this work, we describe the organization and gene structure of a proteasome β-type subunit cluster, comprising an LMP7 pseudogene, LMP2, LMP2-like, MECL-1, and LMP7 in Fugu. We have reanalyzed sequencing and mapping data for these genes in other species and reassess the evolutionary history of this family.

Cosmid 103D19 was identified by hybridization of a probe generated from Fugu clone 161O10cF5 to a 7× Fugu genomic DNA cosmid gridded filter set (available from the HGMP, http://www.hgmp.mrc.ac.uk). This Fugu clone contained an LMP7 fragment (aa 146–174) from an LMP7 pseudogene previously identified on a Fugu cosmid 161O10, which was subjected to full-depth sequencing (Clark, unpublished data). Presence of the full-length LMP7 was confirmed by shotgun sequencing and BLAST searches, detailed below.

The 103D19 cosmid DNA was isolated using the standard alkaline lysis method (19). One microgram of DNA was sonicated to an approximate size of 500 bp, end filled with T4 DNA polymerase, PEG precipitated, and ligated into a phosphatased blunt-ended vector, either SmaI-cut M13 (Amersham, Arlington Heights, IL) or EcoRV-cut pBluescript (Stratagene, La Jolla, CA). After transformation, M13 clone DNAs were isolated using the Qiagen (Chatsworth, CA) 9600 robot and sequenced using Amersham-40 M13 ET dye primers at two-thirds reaction volumes. Sequencing reaction conditions were as described by the manufacturer. pBluescript clone inserts were PCR amplified and sequenced using limiting dilutions of dNTPs and primers and PE Biosystems d-Rhodamine terminator reaction mix (http://menu.hgmp.mrc.ac.uk/ISO9000/BIOLOGY/FUGU/Fugu.shtml).

Sequences were transferred to a UNIX environment and quality clipped using a modified Pregap script (20). Sequences were screened against sequencing (pBluescript, M13) and cosmid (Lawrist4) vectors, and matching regions were masked before further analyses. Contig assembly was performed using Gap4 (21). The consensus sequence of this cosmid fragment has been submitted to the EMBL database with the accession number AJ271723. Coding regions of the proteasome subunits were identified by BLAST similarity searches using BLAST v2.0 (22) against the SWISS-PROT (23), TREMBL (23), and EMBL (24) databases via the HGMP NIX interface (http://www.hgmp.mrc.ac.uk/NIX/). Anomalous intron/exon boundaries were confirmed by RT-PCR of exon-specific primers with Fugu spleen first-strand synthesis cDNA. Searches were conducted for potential noncoding conserved motifs using Dotter (25) (whole proteasome genes) and LALIGN (26) (5′ intergenic sequence for LMP7 and LMP2-like and including the first intron of the neighboring gene for both LMP2 and MECL-1, as these two genes are so close together (98 bp)). Fuzznuc from the EMBOSS suite of program (http://www.sanger.ac.uk/Software/EMBOSS/) was used to search for potential IFN-regulatory factor motifs, as defined previously (18).

Sixty-one sequences were used in the analyses representing the proteasome subunits LMPY/LMP2, LMPX/LMP7, and LMPZ/MECL-1 (see Table I for accession numbers). Five new sequences from Fugu (the putative LMP2 (FrLMP2), LMP2-like (FrLMP2-like), LMP7 (FrLMP7), LMP7 pseudogene (FrLMP7-like), and MECL-1 (FrMECL-1)) genes were added. Multiple alignments were constructed using Clustal X (27). To determine the major groups of paralogous/orthologous genes of the different sequences obtained from GenBank, a phylogenetic tree was constructed based on a Poisson correction distance with a pairwise deletion comparison. After that, the three groups were distinguished by high distance values and supported by high bootstrap proportion (BP) (BP = 99–100%). These three groups corresponded clearly to a particular subunit (LMPY/LMP2, LMPX/LMP7, and LMPZ/MECL-1), and each group was then analyzed separately using a Poisson correction distance and a γ distance of α = 2 for these three genes (28). Robustness of nodes was estimated by bootstrap with 1000 replicates (29). All the trees are rooted on midpoint, excepted for the LMPZ/MECL-1 for which Saccharomyces cerevisiae (Sce) was chosen due to its position as the outgroup in the complete tree.

Table I.

Species name and accession numbers of the sequences used in the analyses

GeneSpecies NameCommon NameAllele/GeneAccession No.
PR2 Saccharomyces cerevisiae Yeast  M96667 
LMPX Geodia cydonium Sponge  X97728 
LMPX Botryllus schlosseri Tunicate  X97729 
LMPX Myxine glutinosa Hagfish  D64054 
LMPX Petromyzon marinus Lamprey  D64055 
LMPX Ginglymostoma cirratum Shark  D64058 
LMPX Danio rerio Zebrafish X1 AF032391 
LMPX Danio rerio Zebrafish X2 AF155578 
LMPX Gallus domesticus Chicken  X57210 
LMPX Gallus gallus Chicken  AB001935 
LMPX Mus musculus Mouse PSMB5 AF060091 
LMPX Homo sapiens Human  D29011 
LMP7 Ginglymostoma cirratum Nurse shark 7a 64057 
LMP7 Ginglymostoma cirratum Nurse shark 7b 64056 
LMP7 Oryzias latipes Medaka fish  D89725 
LMP7 Danio rerio Zebrafish  AF032390 
LMP7 Fugu rubripes Japanese pufferfish New 
LMP7 Fugu rubripes Japanese pufferfish 2a New 
LMP7-a Xenopus laevis African clawed toad 7a D44540 
LMP7-b Xenopus laevis African clawed toad 7b D44549 
LMP7 Sus scrofa Pig  AF059493 
LMP7 Homo sapiens Human 7E1 Z14982 
LMP7 Homo sapiens Human 7E2 Z14982 
LMP7 Rattus norvegicus (sp.) Rat  D10729 
LMP7 Mus musculus Mouse 7b U22032 
LMP7 Mus musculus Mouse 7d U22031 
LMP7 Mus musculus Mouse 7k, s, f U22033 
     
PBA1 Arabidopsis thaliana Mouse-ear cress  Y09364 
LMP2 Nicotania tabacum Tobacco  Y09505 
DAPS1 Dictyostelium discoideum Slime mould  AB007024 
PRE3 Saccharomyces cerevisiae Yeast  X78991 
LMPY Lampetra japonica Lamprey  D87690 
LMPY Onchorynchus mykiss Rainbow trout  AF115539 
LMPY Onchorynchus mykiss Rainbow trout  AAD53036 
LMPY Danio rerio Zebrafish  AF032392 
LMPY Xenopus laevis African clawed toad  D87689 
LMPY Homo sapiens Human  D29012 
LMPY Rattus norvegicus (sp.) Rat  D10754 
LMPY Mus musculus Mouse  NM008946 
LMP2-like Fugu rubripes Japanese pufferfish  New 
LMP2-like Onchorynchus mykiss Rainbow trout  AF115540 
LMP2-like Danio rerio Zebrafish  AF155576 
LMP2 Danio rerio Zebrafish PSMB9a AF155579 
LMP2 Danio rerio Zebrafish PSMB9b AF155580 
LMP2 Danio rerio Zebrafish PSMB9c AF032393 
LMP2 Oryzias latipes Medaka fish  D89724 
LMP2 Onchorynchus mykiss Rainbow trout AF112117 
LMP2 Onchorynchus mykiss Rainbow trout AF115541 
LMP2 Onchorynchus mykiss Rainbow trout  AAD53038 
LMP2 Xenopus laevis African clawed toad  D87687 
LMP2 Homo sapiens Human  U01025 
LMP2 Rattus norvegicus (sp.) Rat  D10757 
LMP2 Mus musculus Mouse  U22919 
PUP1 Saccharomyces cerevisiae Yeast  AB000540 
LMPZ Drosophila melanogaster Fruitfly alpha AF025792 
LMPZ Drosophila melanogaster Fruitfly beta AF025791 
LMPZ Danio rerio Zebrafish PSMB7 AF155581 
LMPZ Homo sapiens Human  NP002790 
LMPZ Mus musculus Mouse  D83585 
LMPZ Ciona intestinalis Ascidian  AJ002142 
MECL1 Danio rerio Zebrafish  AF155577 
MECL1 Fugu rubripes Japanese pufferfish  New 
MECL1 Homo sapiens Human  NP002792 
MECL1 Mus musculus Mouse U77784 
MECL1 Mus musculus Mouse Y10875 
MECL1 Mus musculus Mouse D85561 
GeneSpecies NameCommon NameAllele/GeneAccession No.
PR2 Saccharomyces cerevisiae Yeast  M96667 
LMPX Geodia cydonium Sponge  X97728 
LMPX Botryllus schlosseri Tunicate  X97729 
LMPX Myxine glutinosa Hagfish  D64054 
LMPX Petromyzon marinus Lamprey  D64055 
LMPX Ginglymostoma cirratum Shark  D64058 
LMPX Danio rerio Zebrafish X1 AF032391 
LMPX Danio rerio Zebrafish X2 AF155578 
LMPX Gallus domesticus Chicken  X57210 
LMPX Gallus gallus Chicken  AB001935 
LMPX Mus musculus Mouse PSMB5 AF060091 
LMPX Homo sapiens Human  D29011 
LMP7 Ginglymostoma cirratum Nurse shark 7a 64057 
LMP7 Ginglymostoma cirratum Nurse shark 7b 64056 
LMP7 Oryzias latipes Medaka fish  D89725 
LMP7 Danio rerio Zebrafish  AF032390 
LMP7 Fugu rubripes Japanese pufferfish New 
LMP7 Fugu rubripes Japanese pufferfish 2a New 
LMP7-a Xenopus laevis African clawed toad 7a D44540 
LMP7-b Xenopus laevis African clawed toad 7b D44549 
LMP7 Sus scrofa Pig  AF059493 
LMP7 Homo sapiens Human 7E1 Z14982 
LMP7 Homo sapiens Human 7E2 Z14982 
LMP7 Rattus norvegicus (sp.) Rat  D10729 
LMP7 Mus musculus Mouse 7b U22032 
LMP7 Mus musculus Mouse 7d U22031 
LMP7 Mus musculus Mouse 7k, s, f U22033 
     
PBA1 Arabidopsis thaliana Mouse-ear cress  Y09364 
LMP2 Nicotania tabacum Tobacco  Y09505 
DAPS1 Dictyostelium discoideum Slime mould  AB007024 
PRE3 Saccharomyces cerevisiae Yeast  X78991 
LMPY Lampetra japonica Lamprey  D87690 
LMPY Onchorynchus mykiss Rainbow trout  AF115539 
LMPY Onchorynchus mykiss Rainbow trout  AAD53036 
LMPY Danio rerio Zebrafish  AF032392 
LMPY Xenopus laevis African clawed toad  D87689 
LMPY Homo sapiens Human  D29012 
LMPY Rattus norvegicus (sp.) Rat  D10754 
LMPY Mus musculus Mouse  NM008946 
LMP2-like Fugu rubripes Japanese pufferfish  New 
LMP2-like Onchorynchus mykiss Rainbow trout  AF115540 
LMP2-like Danio rerio Zebrafish  AF155576 
LMP2 Danio rerio Zebrafish PSMB9a AF155579 
LMP2 Danio rerio Zebrafish PSMB9b AF155580 
LMP2 Danio rerio Zebrafish PSMB9c AF032393 
LMP2 Oryzias latipes Medaka fish  D89724 
LMP2 Onchorynchus mykiss Rainbow trout AF112117 
LMP2 Onchorynchus mykiss Rainbow trout AF115541 
LMP2 Onchorynchus mykiss Rainbow trout  AAD53038 
LMP2 Xenopus laevis African clawed toad  D87687 
LMP2 Homo sapiens Human  U01025 
LMP2 Rattus norvegicus (sp.) Rat  D10757 
LMP2 Mus musculus Mouse  U22919 
PUP1 Saccharomyces cerevisiae Yeast  AB000540 
LMPZ Drosophila melanogaster Fruitfly alpha AF025792 
LMPZ Drosophila melanogaster Fruitfly beta AF025791 
LMPZ Danio rerio Zebrafish PSMB7 AF155581 
LMPZ Homo sapiens Human  NP002790 
LMPZ Mus musculus Mouse  D83585 
LMPZ Ciona intestinalis Ascidian  AJ002142 
MECL1 Danio rerio Zebrafish  AF155577 
MECL1 Fugu rubripes Japanese pufferfish  New 
MECL1 Homo sapiens Human  NP002792 
MECL1 Mus musculus Mouse U77784 
MECL1 Mus musculus Mouse Y10875 
MECL1 Mus musculus Mouse D85561 
a

Pseudogenes.

Previously, a Fugu cosmid, 161O10, had been identified that on completion of full-depth sequencing was shown to contain an LMP7 pseudogene (the first 144 aa were absent) (Clark, unpublished data). Further hybridization experiments using this LMP7 gene as a probe identified a nonoverlapping cosmid 103D19. Sequence scanning and database similarity searching of this cosmid revealed several genes related to LMPs. It was unclear from the sequence scan data exactly which LMP genes were present, and so a 22.398-kb fragment containing these LMP genes was subjected to full-depth sequencing and further analysis. This subclone contained genes for LMP2, LMP2-like, MECL-1, and LMP7. It also contained a TAP2B and an MHC class Iα gene (Fig. 1) and mapped within a larger bacterial artifical chromosome contig, comprising the Fugu MHC class I region (M. Clark, unpublished data). Further genomic analysis was conducted on all of the Fugu proteasome genes to identify potential regulatory motifs. Dotter pairwise comparisons of whole genes showed very little homology even at the exon level, even between the closely related LMP2 and LMP2-like duplicated genes. Pattern searches for the IFN-regulatory factor motifs (AAAAGYGAAA) over the whole cosmid subclone did not produce any significant matches. The LALIGN program was more successful, showing regions of homology between the different proteasome genes ranging from 60 to 80% identity over 11- to 170-bp regions. However, it was difficult to precisely define a particular region that was either gene specific or proteasome specific, and it is felt that comparison with another fish species would be required to accurately test these results; however, these data are not currently publicly available.

FIGURE 1.

The β proteasome subunit cluster structure of Fugu.

FIGURE 1.

The β proteasome subunit cluster structure of Fugu.

Close modal

The complete tree (rooted at the midpoint) (Fig. 2 A) shows 27 sequences constituting the LMPX/7group (BP = 100%), 26 sequences constituting the LMPY/2 group (BP = 99%), and 13 sequences constituting the LMPZ/MECL-1 group (BP = 100%). Distances within groups are respectively equal to 0.418 (±0.029), 0.544 (±0.035), and 0.521 (±0.032). Distances between the three clusters are equal to 1.248 (±0.090) for LMPX/Y, 1.450 (±0.093) for LMPX/Z, and 1.328 (±0.086) for LMPY/Z. Each subfamily was then reassessed separately.

FIGURE 2.

A, Bootstrap analyses conducted with 1000 iterations using the 66 sequences (LMPY/LMP2, LMPX/LMP7, and LMPZ/MECL-1) using Neighbor Joining on a matrix of Poisson correction distance and on γ distance (α = 2). The Likelihood ratio test indicates that the two trees are not significantly different in topology, so only the tree based on Poisson correction distance is shown. The tree is rooted on the midpoint corresponding, in this case, to S. cerevisiae as an outgroup. Fru prefixes all the new Fugu genes added to the analyses. B, Bootstrap analyses conducted with 1000 iterations using 27 sequences (LMPX/LMP7) using Neighbor Joining on a matrix of Poisson correction distance and on γ distance (α = 2). The Likelihood ratio test indicates that the two trees are not significantly different in topology, so only the tree based on Poisson correction distance is shown. The tree is rooted on the midpoint corresponding, in this case, to S. cerevisiae as an outgroup. Fru prefixes all the new Fugu genes added to the analyses. C, Bootstrap analyses conducted with 1000 iterations using 26 sequences (LMPY/LMP2) using Neighbor Joining on a matrix of Poisson correction distance and on γ distance (α = 2). The Likelihood ratio test indicates that the two trees are not significantly different in topology, so only the tree based on Poisson correction distance is shown. The tree is rooted on the midpoint corresponding, in this case, to A. thaliana-N. tabacum as an outgroup. Fru prefixes all the new Fugu genes added to the analyses. D, Bootstrap analyses conducted with 1000 iterations using 13 sequences (LMPZ/MECL-1) using Neighbor Joining on a matrix of Poisson correction distance and on γ distance (α = 2). The Likelihood ratio tests indicate that the two trees are not significantly different in topology, so only the tree based on Poisson correction distance is shown. The tree is rooted on S. cerevisiae. The midpoint is indicated. Fru prefixes all the new Fugu genes added to the analyses.

FIGURE 2.

A, Bootstrap analyses conducted with 1000 iterations using the 66 sequences (LMPY/LMP2, LMPX/LMP7, and LMPZ/MECL-1) using Neighbor Joining on a matrix of Poisson correction distance and on γ distance (α = 2). The Likelihood ratio test indicates that the two trees are not significantly different in topology, so only the tree based on Poisson correction distance is shown. The tree is rooted on the midpoint corresponding, in this case, to S. cerevisiae as an outgroup. Fru prefixes all the new Fugu genes added to the analyses. B, Bootstrap analyses conducted with 1000 iterations using 27 sequences (LMPX/LMP7) using Neighbor Joining on a matrix of Poisson correction distance and on γ distance (α = 2). The Likelihood ratio test indicates that the two trees are not significantly different in topology, so only the tree based on Poisson correction distance is shown. The tree is rooted on the midpoint corresponding, in this case, to S. cerevisiae as an outgroup. Fru prefixes all the new Fugu genes added to the analyses. C, Bootstrap analyses conducted with 1000 iterations using 26 sequences (LMPY/LMP2) using Neighbor Joining on a matrix of Poisson correction distance and on γ distance (α = 2). The Likelihood ratio test indicates that the two trees are not significantly different in topology, so only the tree based on Poisson correction distance is shown. The tree is rooted on the midpoint corresponding, in this case, to A. thaliana-N. tabacum as an outgroup. Fru prefixes all the new Fugu genes added to the analyses. D, Bootstrap analyses conducted with 1000 iterations using 13 sequences (LMPZ/MECL-1) using Neighbor Joining on a matrix of Poisson correction distance and on γ distance (α = 2). The Likelihood ratio tests indicate that the two trees are not significantly different in topology, so only the tree based on Poisson correction distance is shown. The tree is rooted on S. cerevisiae. The midpoint is indicated. Fru prefixes all the new Fugu genes added to the analyses.

Close modal

S. cerevisiae (Sce) was used as an outgroup (Fig. 2 B). Monophylogeny of both paralogues was highly supported (99% for LMP7 and 99% for LMPX). The lamprey and hagfish, which contain the precursors of these genes before they duplicated into LMPX and LMP7, constitute a monophyletic group (BP = 96%) and cluster with the LMP7 genes (BP = 80%). The two new Fugu sequences (Fru7 and Fru7-like) cluster with LMP7 sequences from other fish species (Oryzias latipes (Ola; medaka) and Dre (zebrafish)). Furthermore, the Fugu LMP7 and LMP7 psuedogene are monophyletic. Interestingly, the clustering and bootstrap values are very high within the LMP7 group. Indeed, the mammalian LMP7 genes cluster strongly with the amphibian representatives Rattus norvcgicus (Rno) and Xenopus laevis (Xla) (BP = 99%), which cluster with the teleostean fishes (76%) and with the shark species (Ginglymostoma cirratum, Gci) (99%). No obvious clustering relationships were observed for the LMPX gene.

Arabidopsis thaliana (Ath) and Nicotania tabacum (Nta) were used as an outgroup (Fig. 2,C). The tree topology of the LMPY/LMP2 is different to the complete tree (whole data set), suggesting a huge impact of the other sequences (LMPX/LMP7; LMPZ/MECL-1) on the internal topology of the LMPY/LMP2 relationships (cf Fig. 2,A). In the separate analysis presented in Fig. 2 C, S. cerevisiae (Sce) is also found in a basal position with Dictyostelium discoideum (Ddi). Two main clusters are observed, one represented by LMPY (BP = 88%) and the other by the LMP2 and LMP2-like (a duplication of LMP2) genes (BP = 98%). One Fugu (Fru2-like) sequence clusters with the trout (Omy) and zebrafish (Dre) LMP2-like genes (BP = 100%) and the other with the trout (Omy), medaka (Ola), and zebrafish (Dre) LMP2 genes (BP = 100%). The clustering pattern is very tight for both paralogous genes (LMPY and LMP2). Humans (Homo sapiens, Hsa) are clearly shown as the closest relatives of the mouse (Mus musculus, Mmu) and rat (Rno) (BP = 100% in both cases). Xenopus (Xla) clusters with the mammals (BP = 71% and 63%), although this clade is the closest relative of the teleostean fishes (Fugu rubripes (Fru), Omy, Ola, and Dre). The lamprey sequence (LY), which is considered to be the precursor of these genes before they duplicated into LMP2, LMP2-like, and LMPY, clusters with the LMPY sequences in a basal position (BP = 99%).

S. cerevisiae (Sce) was used as an outgroup (Fig. 2 D). The final new Fugu sequence (FrMECL) clusters with zebrafish PSMB12 (DreMECL) (BP = 99%). These two sequences cluster with the human and mouse MECL-1 genes (HsaMECL and MmuMECL1–3) (BP = 80%). The paralogous genes (LMPZ) group with Ciona in a trichotomy (BP = 52%), and so the exact nature of the evolutionary interrelationships is difficult to determine.

All Fugu full-length proteasome genes described in this work (LMP2, LMP2-like, LMP7, and MECL-1) have the same number of exons as their mammalian homologues, with all splice junctions following the GT/AG rule (30). The boundary classes of the introns/exons are perfectly conserved between Fugu and human (31 , accession number: NP002792) for all four genes, with the Fugu LMP2-like gene being compared with the human LMP2. Results from the comparison of exon sizes between Fugu and human mirror the phylogenetic analyses; the sizes of seven of the eight MECL-1 exons are perfectly conserved, the sizes of four of the six LMP7 exons are conserved, and in the case of LMP2, five of the six exon sizes are conserved, with four of the six being conserved when the Fugu LMP-like gene is compared with the human LMP2.

Phylogenetic analyses confirm the presence of the three IFN-γ-inducible β proteasome subunits (FrLMP2, FrLMP7, and FrMECL-1) at a single locus in Fugu. All mammals, anuran amphibians, and fish (16, 18, 32, 33, 34, 35, 36, 37, 38, 39) that have been analyzed, to date, for the presence of the proteasome β subunits have shown conserved linkage of LMP2 and LMP7, both of which reside within the MHC. MECL-1 has, to date, only been mapped in human and mouse, to 16q22.1 (40) and mouse chromosome 8 (8), respectively. Recently, a 26-kb proteasome β subunit cluster has been mapped and partially sequenced in zebrafish. The organization of this cluster mirrors that of Fugu; however, the authors concluded that it contained two unique proteasome genes, one of which they designated PSMB12, belonging to the Z group of proteasome genes (18). Our more extensive phylogenetic analysis reveals that PSMB12 is also present in Fugu and is in fact MECL-1. Due to the differential rate of substitution patterns between the three families (as defined in Materials and Methods), conclusions about the phylogenetic relationships inside the three families can be erroneous if one uses the complete data set (LMPX-7, LMPY-2, LMPZ-MECL-1). A sequence belonging to one family can be misplaced (in its own family) due to the impact of the other sequences belonging to the other two families. This is why some sequences are misplaced in the tree obtained by the using the Neighbor joining method with the complete data set described previously (18). The addition of new sequences and the use of separate analyses for each family is the best way to remove these artifacts.

Additionally within this region in Fugu, there is a duplication of LMP2 to form an LMP2-like gene. To date, this gene has only been identified in zebrafish (18), designated a novel Y proteasome gene (PSMB11), and rainbow trout (17), designated LMP2/δ. In all three teleosts, the LMP2 and LMP2-like sequences are closely linked and in the same transcriptional orientation. Therefore, the LMP2-like gene probably arose due to a cis duplication event from LMP2. Analysis in the rainbow trout reveals that this LMP2-like gene has the same expression profile as LMP2, with highest expression in the trout lymphoid tissues, similar to the class I alleles and consistent with LMP expression in other systems (17). As these two genes both reside within the teleostean MHC and extensive analysis conducted in other vertebrate MHCs has not yet revealed a similar sequence, it is proposed that the cis duplication of LMP2 is specific, at least to the teleosts. As it has an expression pattern that mirrors the original LMP2, it is almost certainly involved in the class I Ag presentation and may contribute to Ag presentation diversity.

Closely linked to the proteasome β subunit cluster in Fugu (within 100 kb), there is an LMP7 pseudogene. This is clearly nonfunctional, as the first 144 aa are missing. To date, this pseudogene has not been identified in other fish species. This may be for either of two reasons: the LMP7 duplication is Fugu specific or, because the main method of gene identification and mapping in other species has been using gene-specific primers, the design of these may not allow the amplification of such a 3′ gene fragment. Without more extensive analyses of the MHC in other fishes, it will not be possible to determine which of these two scenarios is correct. To date, there is no evidence in Fugu for a second MECL-1 gene, so the most probable explanation for the presence of the LMP7 pseudogene is a second independent cis duplication event. This is supported by phylogenetic analysis, which indicates that the LMP7/LMP7 pseudogene duplication occurred very late in the Fugu lineage, as the sequences are almost identical. The LMP2 duplication occurred much earlier and therefore represents an independent evolutionary event. The entire Fugu MHC class I region contains a number of different transposon species (M. S. Clark, unpublished data), and retroelement action has been implicated in the evolution of the MHC region, both globally (41, 42) and also in a more specific way with transposition, gene structure organization, and evolution of processed pseudogenes (43, 44). The second Fugu LMP7 may have been translocated and mutated via transposon action to a region on the same chromosome, still within the MHC class I region.

This Fugu cosmid also contains a TAP2 gene that is orthologous to the trout TAP2B gene (17). The current position of the Fugu TAP2A is not known. Teleosts have two types of TAP2, TAP2A and TAP2B, indicating that once again, more duplications have occurred in the teleost lineage compared with other vertebrates, specifically the mammals, in which extensive data exist. In trout, TAP2B is linked to LMP2, LMP2/δ, and the MHC class Iα region (17); TAP2A has been identified, but is, as yet, unmapped. The zebrafish TAP2 has not been accurately defined, but it would appear, on the current data available, that the consensus teleost linkage is TAP2B, LMP7, LMP2, and MECL-1. This Fugu cosmid, in addition to the TAP2B gene, also contains an MHC class Iα gene and maps within a larger MHC class I bacterial artifical chromosome contig (M. S. Clark, unpublished data), thus confirming linkage of LMP2 and LMP7 to the MHC class I region over 420 million years, which has been demonstrated in other fish species and amphibians (16, 18, 35).

The processes behind the evolutionary development of the MHC are controversial. There are currently two main hypotheses. First, the linkage relationships in the MHC are the result of adaptive evolution. Genes within gene families (e.g., ABC transporters, LMPs, etc.) duplicated at different times in evolution, and genes with a similar function and expression pattern were recruited over a long evolutionary timescale to the same chromosomal region. This includes the genes that are involved in the Ag presentation pathway (MHC class Iα, TAP, LMPs), which remain linked due to the advantage of expression coregulation by IFN-γ (45) (Fig. 3,A). Second, the ancestors of genes within the MHC, such as the β-type proteasome subunits, heat-shock protein 70, and members of the ABC transporter superfamily, were originally recruited to a single chromosomal region, at or before the emergence of the vertebrates. This region underwent a duplication event (either chromosome or whole genome) at an early stage of vertebrate evolution, with one of the duplicated regions forming the ancestral MHC (6p21.3). Genes within these newly formed duplicated regions then either remained linked due to common functionality or translocated elsewhere in the genome (13) (Fig. 3 B).

FIGURE 3.

Representations of the two main possibilities for β-type proteasome subunit organization in vertebrates. A, Independent duplication events in the vertebrate ancestor produced the six β-type proteasome subunits (X, Y, Z, X′, Y′, and Z′), which were not linked. In mammals, the LMP7 and LMP2 subunits were recruited into the MHC class II region due to their coordinated role in Ag presentation and regulation by IFN-α. In teleosts, LMP7, LMP2, and MECL-1 were all independently recruited to the class I region for the same reasons. B, The ancestral X/Y/Z-like β-type proteasome subunits were linked in the eukaryote ancestor. A large-scale or whole genome duplication event occurred producing two separate loci containing X, Y, Z and X′, Y′, and Z′. The original X, Y, and Z gene cluster fragmented, and the genes were translocated throughout the genome. The X′, Y′, and Z′ gene cluster evolved into LMP2, LMP7, and MECL-1. These remained linked in the teleost ancestor to the MHC class I region, while in mammals, LMP2 and LMP7 remained in the ancestral MHC, with MECL-1 being translocated to another chromosome.

FIGURE 3.

Representations of the two main possibilities for β-type proteasome subunit organization in vertebrates. A, Independent duplication events in the vertebrate ancestor produced the six β-type proteasome subunits (X, Y, Z, X′, Y′, and Z′), which were not linked. In mammals, the LMP7 and LMP2 subunits were recruited into the MHC class II region due to their coordinated role in Ag presentation and regulation by IFN-α. In teleosts, LMP7, LMP2, and MECL-1 were all independently recruited to the class I region for the same reasons. B, The ancestral X/Y/Z-like β-type proteasome subunits were linked in the eukaryote ancestor. A large-scale or whole genome duplication event occurred producing two separate loci containing X, Y, Z and X′, Y′, and Z′. The original X, Y, and Z gene cluster fragmented, and the genes were translocated throughout the genome. The X′, Y′, and Z′ gene cluster evolved into LMP2, LMP7, and MECL-1. These remained linked in the teleost ancestor to the MHC class I region, while in mammals, LMP2 and LMP7 remained in the ancestral MHC, with MECL-1 being translocated to another chromosome.

Close modal

We believe that the organization of the β-type proteasome subunits in Fugu supports the latter theory, that these genes arose from an ancestral X-, Y-, and Z- like gene cluster followed by a large-scale genome duplication event, before the divergence of the jawless and jawed vertebrates, ∼600 million years ago (Fig. 3 B). After this duplication event, the arrangement of X, Y, Z gene cluster became fragmented. In humans, the Z subunit remained on one of the ancestral duplicated regions (9q34), with the other genes translocating to 14q11.2 (X) and 17p13 (Y). The X′, Y′, and Z′ genes evolved into the current LMP7, LMP2, and MECL-1. In Fugu and other fish species, they remained as one cluster linked to the MHC class I region, the linkage maintained by their mutual regulation by IFN-γ and function in the Ag-presenting pathway. In humans, the LMP7 and LMP2 genes remained linked, with MECL-1 translocating to 16q22.1. Therefore, it is proposed that Fugu represents the ancestral arrangement of these sequences in vertebrates.

Whether the situation in Fugu describes the proto-MHC concerned with class I Ag presentation or is a derived condition awaits further evidence from more primitive classes such as the cartilaginous fishes (46). What is certain is that the evolution of the MHC region, in any species, is not simple to explain. While these large-scale duplications were major events, within the evolution of MHC itself (6p21.3) and indeed, in the other paralogous regions (9q34, 1q21–25, and 19p13.1–13.3), many smaller scale rearrangements have taken place involving gene shuffling, cis duplication, translocation, and deletion (13, 14, 15, 47). Comprehensive phylogenetic analysis indicate that some of the genes present in the paralogous regions resulted from direct duplications (e.g., retinoic acid receptors, complement genes, etc.) before the vertebrate radiation, but after the deuterosome radiation, while others (e.g., heat-shock protein 70 family) occurred before the eukaryote radiation (48, 49, 50). The three IFN-γ-regulated β-type proteasome subunits described in this study are themselves a case in point, with evidence of a tandem duplication before the divergence of animals from fungi and the proposed large-scale duplication events (50) and with, it would now appear, at least one further cis duplication present in bony fishes.

The organization found in Fugu and zebrafish could be explained in two ways. First, that X, Y, and Z were not linked in our vertebrate ancestor and after duplication (which may or may not have been by large-scale duplication), X′, Y′, and Z′ were recruited to the same region due to positive selective forces. The question is, when did this occur? Were X′ and Y′ linked before the bony vertebrate radiation, with the linking of Z′ to the two others occurring either only in the teleost lineage or in the bony vertebrate lineage, followed by genomic rearrangement in the mammalian lineage? Second, X′, Y′, and Z′ were linked in our vertebrate ancestor and remain linked in teleosts, but became rearranged in the mammalian lineage. The mapping of MECL-1 in other tetrapod species and in cartilaginous vertebrates (e.g., sharks) will answer such a question.

Further data, particularly the mapping of MECL-1 in other tetrapod species and in cartilaginous vertebrates (e.g., sharks), will enable us to more accurately answer these questions.

1

This work was supported by a Medical Research Council Project Grant.

3

Abbreviations used in this paper: LMP, low molecular mass polypeptide; BP, bootstrap proportion; MECL, multicatalytic endopeptidase complex subunit MECL-1; PSMB, proteasome subunit component β.

1
Monaco, J. J., D. Nandi.
1995
. The genetics of proteasomes and antigen processing.
Annu. Rev. Genet.
29
:
729
2
Monaco, J. J., H. O. McDevitt.
1984
. H-2-linked low-molecular weight polypeptide antigens assemble into an unusual macromolecular complex.
Nature
309
:
797
3
Monaco, J. J., H. O. McDevitt.
1986
. The LMP antigen: a stable MHC controlled multisubunit protein complex.
Hum. Immunol.
15
:
416
4
Martinez, C. K., J. J. Monaco.
1991
. Homology of proteasome subunits to a major histocompatibility complex-linked LMP genes.
Nature
353
:
664
5
Kelly, A., S. H. Powis, R. Glynne, E. Radley, S. Beck, J. Trowsdale.
1991
. Second proteasome-related gene in the human MHC class II region.
Nature
353
:
667
6
Fruh, K., Y. Yang, D. Arnold, J. Chambers, L. Wu, J. B. Waters, T. Spies, P. A. Peterson.
1992
. Alternative exon usage and processing of the major histocompatibility complex-encoded proteasome subunits.
J. Biol. Chem.
267
:
22131
7
Hisamatsu, H., N. Shimbara, Y. Saito, P. Kristensen, K. B. Hendil, T. Fujiwara, E.-I. Takahashi, N. Tanahashi, T. Tamura, A. Ichihara, K. Tanaka.
1996
. Newly identified pair of proteasome subunits regulated reciprocally by interferon γ.
J. Exp. Med.
183
:
1807
8
Hayashi, M., T. Ishibashi, K. Tanaka, M. Kasahara.
1997
. The mouse genes encoding the third pair of β-type proteasome subunits regulated reciprocally by IFN-γ.
J. Immunol.
159
:
2760
9
Akiyama, K., K. Yokota, S. Kagawa, N. Shimbara, T. Tamura, H. Akioka, H. G. Nothwang, C. Noda, K. Tanaka, A. Ichihara.
1994
. cDNA cloning and interferon γ down-regulation of proteasomal subunits X and Y.
Science
265
:
1231
10
Fruh, K., K. Gossen, H. Wang, H. Bujard, P. A. Peterson, Y. Yang.
1994
. Displacement of housekeeping proteasome subunits by MHC-encoded LMPs: a newly discovered mechanism for modulating the multicatalytic proteinase complex.
EMBO J.
13
:
3236
11
Driscoll, J., M. G. Brown, D. Finley, J. J. Monaco.
1993
. MHC-linked LMP gene products specifically alter peptidase activities of the proteasome.
Nature
356
:
262
12
Gaczynska, M., K. L. Rock, A. L. Goldberg.
1993
. γ Interferon and expression of MHC genes regulate peptide hydrolysis by proteasomes.
Nature
365
:
264
13
Kasahara, M., M. Hayashi, K. Tanaka, H. Inoko, K. Sugay, T. Ikemura, T. Ishibashi.
1996
. Chromosomal localization of the proteasome Z subunit reveals an ancient chromosomal duplication involving the major histocompatibility complex.
Proc. Acad. Natl. Sci. USA
93
:
9096
14
Kasahara, M..
1999
. Genome dynamics of the major histocompatibility complex: insights from genome paralogy.
Immunogenetics
50
:
134
15
Kasahara, M..
1999
. The chromosomal duplication model of the major histocompatibility complex.
Immunol. Rev.
167
:
17
16
Namikawa-Yamada, C., K. Naruse, H. Wada, A. Shima, N. Kuroda, M. Nonaka, M. Sasaki.
1997
. Genetic linkage between the LMP2 and LMP7 genes in the medaka fish, a teleost.
Immunogenetics
46
:
431
17
Hansen, J. D., P. Strassburger, G. H. Thorgaard, W. P. Young, L. Du Pasquier.
1999
. Expression, linkage and polymorphism of MHC-related genes in rainbow trout, Onchorhynchus mykiss.
J. Immunol.
163
:
774
18
Murray, B. W., H. Sultman, J. Klein.
1999
. Analysis of a 26kb region linked to the MHC in zebrafish: genomic organization of the proteasome component β/transporter associated with antigen processing-2 gene cluster and identification of five new proteasome β subunit genes.
J. Immunol.
163
:
2657
19
Birmboim, H. C., J. Doly.
1979
. A rapid alkaline extraction procedure for screening recombinant DNA.
Nucleic Acids Res.
7
:
1513
20
Bonfield, J. K., R. Staden.
1996
. Experiment files and their application during large-scale sequencing projects.
DNA Sequence
6
:
109
21
Bonfield, J. K., K. F. Smith, R. Staden.
1995
. A new sequence assembly program.
Nucleic Acids Res.
24
:
4992
22
Altschul, S. F., T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller, D. J. Lipman.
1997
. Gapped BLAST and PSI-BLAST: a new generation of prtoein database search programs.
Nucleic Acids Res.
25
:
3389
23
Bairoch, A., R. Apweiler.
1997
. The SWISS-PROT protein sequence data bank and its supplement TREMBL.
Nucleic Acids Res.
25
:
31
24
Rodriguez-Tomi, P., P. J. Stoehr, G. N. Cameron, T. P. Flores.
1996
. The European Bioinformatics Institute (EBI) databases.
Nucleic Acids Res.
24
:
6
25
Sonnhammer, E. L. L., and R. Durbin. 1995. A dot matrix program with dynamic threshold control suited for genomic DNA and protein alignment. Gene. 167:GC1.
26
Huang, X., W. Miller.
1991
. A time efficient, linear-space local similarity algorithm.
Adv. Appl. Math.
12
:
373
27
Higgins, D. G., A. J. Bleasby, R. Fuchs.
1991
. CLUSTAL V: improved software for multiple sequence alignment.
CABIOS
8
:
189
28
Wang, D. Y.-C., S. Kumar, B. Hedges.
1999
. Divergence time estimates for the early history of animal phyla and the origin of plants, animals and fungi.
Proc. R. Soc. London Ser. B
266
:
163
29
Felsenstein, J..
1985
. Confidence limits on phylogenies: an approach using the bootstrap.
Evolution
39
:
783
30
Brunak, S., J. Engelbrecht, S. Knudson.
1991
. Prediction of human mRNA donor and acceptor sites from the DNA sequence.
J. Mol. Biol.
220
:
49
31
Beck, S., A. Kelly, E. Radley, F. Khurshid, R. P. Alderton, J. Trowsdale.
1992
. DNA sequence analysis of 66kb of the human MHC class II region encoding a cluster of genes for antigen processing.
J. Mol. Biol.
228
:
433
32
Campbell, R. D., J. Trowsdale.
1993
. Map of the human MHC.
Immunol. Today
14
:
349
33
Nonaka, M., C. Namikawa-Yamada, M. Sasaki, L. Salter-Cid, M. F. Flajnik.
1997
. Evolution of proteasome subunits γ and LMP2.
J. Immunol.
159
:
734
34
Namikawa, C., L. Salter-Cid, M. F. Flajnik, Y. Kato, M. Nonaka, M. Sasaki.
1995
. Isolation of Xenopus LMP-7 homologues: striking allelic diversity and linkage to MHC.
J. Immunol.
155
:
1964
35
Takami, K., Z. Zaleska-Rutczynska, F. Figueroa, J. Klein.
1997
. Linkage of LMP, TAP, and RING3 with MHC class I rather than class II genes in the zebrafish.
J. Immunol.
159
:
6052
36
Chardon, P., C. Renard, M. Vaiman.
1999
. The major histocompatibility complex in swine.
Immunol. Rev.
167
:
179
37
O’Brien, S. J., N. Yuhki.
1999
. Comparative genome organization of the major histocompatibility complex: lessons from the Felidae.
Immunol. Rev.
167
:
133
38
Ohta, Y., S. J. Powis, W. J. Coadwell, D. E. Haliniewski, Y. Liu, H. Li, M. F. Flajnik.
1999
. Identification and genetic mapping of Xenopus TAP2 genes.
Immunogenetics
49
:
171
39
Lewin, H. A..
1996
. Genetic organization, polymorphism and function of the bovine MHC. L. B. Schook, and S. J. Lamont, eds.
The Major Histocompatibility Complex of Domestic Animal Species
65
-98. CRC Press, Boca Raton.
40
Larsen, F., J. Solheim, T. Kristensen, A.-B. Kolsto, H. Prydz.
1993
. A tight cluster of five unrelated human genes on chromosome 16q22.1.
Hum. Mol. Genet.
2
:
1589
41
Dawkins, R. L., C. Leelayuwat, S. Gaudieri, G. Tay, J. Hui, S. Cattley, P. Martinez, J. Kulski.
1999
. Genomics of the major histocompatibility complex: haplotypes, duplication, retroviruses and disease.
Immunol. Rev.
167
:
275
42
Kulski, J. K., S. Gaudieri, M. Bellgard, L. Balmer, K. Giles, H. Inoko, R. L. Dawkins.
1997
. The evolution of MHC diversity by segmental duplication and transposition of retroelements.
J. Mol. Evol.
45
:
599
43
Abdulla, S., S. Beck, M. Belich, A. Jackson, T. Nakamura, J. Trowsdale.
1996
. Divergent intron arrangement in the MB1/LMP7 proteasome gene pair.
Immunogenetics
44
:
254
44
Kohda, K., Y. Matsuda, T. Ishibashi, K. Tanaka, M. Kasahara.
1997
. Structural analysis and chromosomal localization of the mouse Psmb5 gene coding for the constitutively expressed β-type proteasome subunit.
Immunogenetics
47
:
77
45
Yeager, M., A. L. Hughes.
1999
. Evolution of the mammalian MHC: natural selection, recombination, and convergent evolution.
Immunol. Rev.
167
:
45
46
Flajnik, M. F., Y. Ohta, C. Namikawa-Yamada, M. Nonaka.
1999
. Insight into the primordial MHC from studies in ectothermic vertebrates.
Immunol. Rev.
167
:
59
47
Abi Rached, L., M. F. McDermott, P. Pontarotti.
1999
. The MHC big bang.
Immunol. Rev.
167
:
33
48
Endo, T., T. Imanishi, T. Gojobori, H. Inoko.
1997
. Evolutionary significance of intragenome duplications on human chromosomes.
Gene
205
:
19
49
Hughes, A. L..
1998
. Phylogenetic tests of the hypothesis of block duplication of homologous genes on human chromosomes 6, 9 and 1.
Mol. Biol. Evol.
15
:
854
50
Hughes, A. L..
1997
. Evolution of the proteasome components.
Immunogenetics
46
:
82