| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Department of Microbiology and Immunology, University of Maryland School of Medicine, Baltimore, MD 21201;
Department of Microbiology and Immunology, University of Mississippi School of Medicine, Jackson, MS 39216; and
Department of Microbiology and Immunology, University of Miami School of Medicine, Miami, FL 33101
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, three components of the constitutive catalytic proteasome
core, X/PSMB5/
5, Y/
/PSMB6/1, and Z/PSMB7/2, are replaced by
LMP7/PSMB8/5i, LMP2/PSMB9/1i, and MECL-1/PSMB10/2i,
respectively, to produce immunoproteasomes, which are believed to
generate peptides suitable in length and sequence for binding to class
I molecules (2, 3, 4). Peptides transported into the
endoplasmic reticulum via a heterodimer of TAP proteins are loaded onto
class I molecules, which then are competent to be expressed at the cell
surface (5). In eukaryotes, one of the first examples of genetic linkage between functionally related, yet structurally unrelated loci was shown for the class I Ag pathway genes, LMP2, LMP7, TAP1, and TAP2, which are closely linked and encoded in the MHC of all vertebrates studied to date (6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17). In mammals, these genes are located in the class II region, while both classical and nonclassical class I genes are encoded in another region of the MHC. The MHC organization of chickens (11) and all teleost fish (9, 13, 14, 15, 16) examined is quite different from mouse and human: classical class I genes are closely linked to proteasome and transporter genes (except in chicken, the inducible proteasome elements have been deleted from the MHC and perhaps from the genome entirely). Such close linkage among class I and the functionally related proteasome and transporter genes has been proposed to be a true class I region, which may allow coevolution of these components (9, 10, 11). The class I region in teleosts is further highlighted because class II genes are found in another region of the genome entirely, so that there are actually (at least) two MHCs in the bony fish (14, 18). Because class I is closely linked to proteasome/transporter genes in all nonmammalian species studied to date, it is likely that this MHC organization is primitive.
In addition to differences in MHC genomic organization in birds and teleosts, suggesting coevolution of class I pathway genes, there are lineages of TAP alleles that have coevolved to function with sets of class I alleles in the rat (19), hamster (20), and probably the chicken (11, 21). In the frog, Xenopus allelic lineages of TAP and LMP7 are linked to particular class I lineages (22, 23, 24). Coevolution of the proteases, transporters, and presenting molecules in the class I Ag presentation pathway may be the selecting force to preserve linkage of the genes in many vertebrate species. Consistent with this idea is the relatively low number of class I genes that are linked to the MHC proper in most nonmammalian vertebrates (excluding the cichlid (25, 26), the axolotl (27, 28), and perhaps the cod (29)), i.e., a stable class I region.
Cartilaginous fish (sharks, skates, rays, and chimeras) are derived from the oldest vertebrate ancestor with an adaptive immune system, and as such are important models for study of the MHC (7). Previously, we have shown that, in contrast to teleost fish, class I and class II genes are closely linked in sharks (1). In addition, previous work demonstrated the existence of TAP1, TAP2 (30), LMP7 homologues (31), and a proteasome gene unique to the cartilaginous fish, called LMP7-like (31). Based on differences in their catalytic sites, it was suggested that proteasomes containing LMP7 or LMP7-like might provide a diverse repertoire of peptides for class I molecules, consistent with the idea that certain class I alleles could bind peptides preferentially produced by one or the other enzyme. In this study, we begin to focus on whether the class I region found in all teleosts and chickens (and probably amphibians) exists in cartilaginous fish by investigating the linkage between class I and class I-related genes (LMP2, LMP7, LMP7-like, TAP1, and TAP2). We also analyze the number of class I genes and their linkage to each other and to other MHC-encoded genes.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A pregnant nurse shark (Ginglymostoma cirratum, order Orectolobiformes) was captured off Little Torch Key, FL, in October 1998, and 39 pups were delivered by Caesarean section (1). Nurse sharks are ovoviviparous, i.e., maintain their young within the uterus, where they are nourished by egg yolk; the size of the yolk sac was an indication of the pups’ ages (32). Shark Yellow used for preparation of the cosmid library is maintained in an outdoor facility in Miami, Florida, with running seawater.
LMP2 isolation
PCR via 5' RACE with a degenerate primer was performed to
isolate the LMP2 gene. The primer
5'-GCRTCNACRTWNCCRTADAT-3', in which R is A or G, N is A or G or C or
T, W is A or T, D is G or A or T (aa position 135–141
(33), Fig. 1
A), and the T3 primer on the
Uni-ZAP phage vector (Stratagene, La Jolla, CA) were used with a horn
shark (Heterodontus francisci) spleen cDNA library as
template (34). PCR was performed for 4 min at 94°C,
followed by 35 cycles of 94°C for 30 s, 45°C for 1 min, 72°C
for 1 min, and a final extension at 72°C for 15 min. A single 600-bp
fragment was amplified, cloned into the pCR2.1 TA cloning vector
(Invitrogen, San Diego, CA), and sequenced.
|
The 600-bp LMP2 PCR fragment was used to screen the
horn shark spleen cDNA library. This horn shark library was also
screened with the full-length nurse shark LMP7 probe
(accession D64057). A full-length LMP2 insert was used to
probe a nurse shark spleen cDNA library prepared from siblings 34 and
37 (35). This nurse shark spleen cDNA library was also
screened with horn shark TAP2 (AF108387, AF108386) and
TAP1 (AF108385) (30) to obtain the orthologous
nurse shark cDNAs, and with nurse shark nonclassical class I
(UAA-NC1, 1) using a probe encoding the 
1 and 2
domains (AF220360) to obtain full-length clones. All library screenings
were done under high stringency conditions (34).
Sequence alignment and phylogenetic tree
The mature LMP2 and LMP7 protein sequences were aligned using Clustal W with minor adjustments, and an unrooted tree was made using the TreeView 1.5 program (36).
Southern blotting
A total of 1–3 ml blood was collected from each pup from the
caudal vein, and DNA was prepared from nucleated erythrocytes, as
described previously (37). Genomic DNA (5 µg) was
digested with various restriction enzymes
(RE)3 to obtain useful
restriction fragment-length polymorphism (RFLP). Full-length nurse
shark cDNA inserts were used as probes, except for TAP2 (it
lacked the region encoding the N-terminal 160 aa (30)) and
UAA-NC1 (the 1 domain exon was used), and all
hybridizations were performed under high stringency conditions.
Cosmid library screening
The nurse shark erythrocyte genomic DNA cosmid library prepared
from shark Yellow was custom-made (SuperCos1 cosmid library;
Stratagene) and has average insert sizes of 
40 kb. Membranes from
library lifts were prepared per the manufacturer’s recommendation with
minor modifications (38), and 800,000 clones were
screened with the full-length nurse shark LMP7 (D64057
(31)) or class Ia probes (UAA01: AF220063
(1)) under high stringency conditions.
Restriction mapping and subcloning
The cosmid clone 32 was digested with BamHI,
SacI, ScaI, XbaI, and all possible
combinations of two enzymes, and ClaI, SmaI,
BamHI, and ScaI were used for digestion of
cosmids 19 and 65. Gels were stained with ethidium bromide to visualize
DNA fragments and then blotted onto nitrocellulose membranes. Membranes
were sequentially hybridized with either
32P-labeled double-strand probes (labeled with
the random-primed labeling method; Boehringer Mannheim, Indianapolis,
IN) or end-labeled exon-specific oligonucleotides to identify the
digested fragments. Cosmid DNAs were also partially digested with RE
and hybridized with end-labeled T3 or T7 oligomers to construct
restriction maps. The LMP7-positive
BamHI-XbaI 4.5-kb fragment (cosmid 32) and the
ClaI-SmaI 1.9-kb fragment (cosmid 65) were
subcloned into the pBluescript cloning vector and sequenced to identify
the exons. For comparison with hybridizing genomic DNA fragments, 10
µg of cosmid DNA was digested with 15 U of RE for 3 h, and
digests were checked on a 1% agarose gel and diluted to convenient
concentrations; 50 ng of cosmid DNA was loaded onto the gel compared
with 5 µg of genomic DNA. A class I 1 domain-specific primer
(5'-TCTCACAGTCTCCGGTATTT-3') and a highly conserved 3 domain primer
(5'-CCATCGTGATCCACCACACA-3') were used to amplify the class I sequence
from the cosmid 32 DNA; PCR was performed for 4 min at 94°C, followed
by 35 cycles of 94°C for 1 min, 56°C for 1 min, 72°C for 3 min,
and a final extension at 72°C for 15 min. A 1.5-kb fragment was
amplified, cloned into pCR2.1 TA cloning vector, and sequenced.
PCR of genomic DNA
To obtain the transmembrane/cytoplasmic region genomic sequence for the UAA-NC1 gene, a UAA/UAA-NC1 exon 5 primer, 5'-TGAACCAAAGACCAGCTC-3', and a UAA-NC1-specific exon 6 primer, 5'-TGGATTGTAGCCAGTCTC-3', were used. PCR was performed at 94°C for 4 min, followed by 35 cycles of 94°C for 1 min, 52°C for 1 min, and 72°C for 3 min, with final extension of 72°C for 15 min using 2–500 ng "mother" genomic DNA as template. A 1-kb product was cloned into pCR2.1 TA cloning vector (Invitrogen) and sequenced. To obtain the UAA transmembrane/cytoplasmic genomic sequence, PCR was performed using cosmid 65 DNA as template. The same exon 5 primer was used as the forward primer with a UAA-specific reverse primer: 5'-AAGCTTAAATCAGCAGGT-3'. PCR was performed as above. The 2.5-kb product was cloned into the pCR2.1 TA cloning vector and sequenced. The PCR-amplified fragment for UAA exon 5 and intron 5 of cosmid 19 was also sequenced.
Northern blotting
RNA was isolated from various tissues using the TRIzol reagent
(Life Technologies, Rockville, MD). Total RNA (20 µg) was
electrophoresed in 1% agarose gels using MOPS as running buffer
(Quality Biologicals, Gaithersburg, MD), and then transferred to an
Optitran nitrocellulose membrane (Schleicher & Schuell, Keene, NH). The
membranes were hybridized with UAA 3' untranslated region
(UTR), UAA-NC1 (1 domain and intron 2), and nucleoside
diphosphate kinase probes (39) on the tissue distribution
blot. The 3' UTR of LMP7 and LMP7-like were used
as probes with RNA made from PBL of 8 sibling sharks, shark
Yellow, and 11 wild animals. All blotting was done under high
stringency conditions (34).
RT-PCR
cDNA was prepared from 5 µg of total RNA from shark Yellow PBL
with the Superscript preamplification system for first-strand cDNA
synthesis kit (Life Technologies). PCR was then performed by 3' RACE
using a UAA 1 domain forward primer,
5'-TCTCACAGTCTCCGGTATTT-3', and the RACE reverse primer,
5'-CTCGAGAAGCTTGAATTCGGATCC-3' at 94°C for 4 min, followed by 35
cycles of 94°C for 1 min, 55°C for 1 min, 72°C for 4 min, and
72°C for 15-min final extension. A 1.7-kb single band was cloned into
pCR2.1 TA cloning vector (Invitrogen) and sequenced.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A 600-bp PCR product was amplified by 5' RACE from a horn shark
spleen cDNA library with a primer specific for LMP2 genes
(33). The sequence of this product showed high similarity
to all known LMP2 genes, and it was used to screen the same
library to isolate LMP2 cDNA clones. The longest horn shark
clone (1044 bp, with a 654-bp open reading frame, 304-bp 5' UTR, and
86-bp 3' UTR, Fig. 1
A) was
then used to screen a nurse shark spleen cDNA library. One full-length
clone was sequenced and consisted of a 119-bp 5' UTR, 61-bp 3' UTR, and
654-bp open reading frame encoding 217 aa (Fig. 1A). The
horn and nurse shark clones are 81% identical at the nucleotide level,
and the deduced mature nurse shark protein sequence (start at +1 in
Fig. 1A) is 87% identical to horn shark LMP2, 67–73% to
LMP2 of other species, and 54–62% to proteasome ( throughout
the text). Nurse and horn shark LMP2 group with other species’ LMP2,
whereas LMP2/ (an intermediate found only in teleosts (14, 15, 40, 41)) and form separate clusters (Fig. 1B).
The nurse shark LMP7-like gene was described previously
(31). To determine whether this gene is expressed by other
elasmobranchs, we screened a horn shark spleen cDNA library and
isolated the orthologous LMP7-like. Rather like the teleost
LMP2/, the sequence of the resulting horn shark clone forms a
cluster with nurse shark LMP7-like in phylogenetic trees, which is
equidistant from the LMP7 and X clusters (Fig. 1C). This
position of the LMP7-like cluster in the tree suggests that the gene
arose not long after the divergence of LMP7 and
X, and indeed it may be present in all extant
cartilaginous fish.
Linkage studies of TAP, LMP, and class I
In the family of 39 nurse sharks, 13 segregating groups (a–m)
were identified previously from classical class I (UAA) RFLP
typing (1). Now we have extended the linkage analysis for
LMP2, LMP7 (31),
LMP7-like (31), TAP1, and
TAP2 (30). Representative Southern blots of 20
of these siblings are shown in Fig. 2, and summaries of all experiments are shown in Tables I and II. TAP1 provided
limited RFLP and formed only three groups (denoted a–c under the
TAP1 blot in Fig. 2, a = U/L, b = L, c = U; U
and L designate the upper and lower bands, respectively, in Tables I and II), but each group matched perfectly
to the class I groups. RFLP for LMP2, LMP7
(EcoRI), and TAP2 were found only for paternal
MHC haplotypes within groups d, e, g, and h (marked with dashes at the
side of each blot, and see Table II). One weakly hybridizing LMP7 band was shown not to be MHC linked (BamHI digest in Fig. 2).
|
|
|
and 
).
The three groups of bands
(a = U/L, b = U, c = L, shown on the bottom of the blot)
matched the class I MHC groups perfectly, and the presence vs absence
of the set
of bands precisely matched paternal haplotypes in groups d, e, g, and
h; thus, both and LMP7-like genes are MHC linked, but only one set of bands is found in all
haplotypes (and see later).
RFLP of all 39 siblings are summarized in Table I, in which
results with the maternal UAA class I 1 domain are shown
to illustrate the inheritance of maternal haplotypes identified
definitively from siblings having distinctive paternal bands
(1). The predicted maternal haplotype m1 was in 18
offspring and is L (UAA), U (TAP1), and U
(LMP7-like );
m2 was in 21 offspring and is U (UAA), L (TAP1),
and L (LMP7-like );
neither maternal haplotype contains the LMP7-like set of bands. Unfortunately, no LMP2, LMP7, or TAP2 RFLP were detected for the maternal haplotypes;
nevertheless, as mentioned above, several paternal haplotypes had
distinctive RFLP mapping to MHC for all of these genes.
Sorting out paternal haplotypes
Twelve paternal haplotypes were detected (p1 through p12), demonstrating that there were at least six fathers of this family
(Table II). Besides the MHC-linked genes, we used the non-MHC-linked
LMP7 RFLP (Fig. 2) and other nonlinked genes (data not
shown), as well as body and yolk sac size, to determine haplotype pairs
for each father. The p1 and p2 haplotypes had RFLP for the nonlinked
LMP7, strongly suggesting that they were derived from one
father (designated father 1, responsible for 17 of 39 offspring). In
addition, siblings belonging to the four groups, a, f, d, and h, were
younger than the others, having larger yolk sacs and smaller body
sizes. Unfortunately, no other haplotypes produced RFLP, or yolk sac
and body size were not informative enough to discriminate the pairs
(the e and g groups had the same paternal RFLP for all genes tested;
however, yolk sac and body length in these two groups are distinct,
suggesting different fathers, or the same father after a long
refractory period). Interestingly, there is an allelic preference for
father 1 since 76% (13 of 17) of the offspring have the p2 haplotype.
We observed the same phenomenon in another family of 18 nurse shark
pups: one haplotype was found in six siblings, and the other haplotype
in a maximum of three pups (1).
Cosmid clones bearing class I and proteasome genes
A genomic cosmid library was prepared from shark Yellow to
study physical linkage of the shark MHC genes. Three class I-positive
clones (cosmids 19, 32, 65) were isolated that hybridized weakly to the
full-length LMP7 probe, with cosmid 32 displaying the
highest intensity signal. Comparison of BamHI digests of
cosmid 32 and shark Yellow genomic DNA resulted in comigration of the
strongly hybridizing band (the weakly hybridizing band is due to
cross-hybridization to the vector) in the cosmid clone and the weakly
hybridizing, non-MHC-linked band in genomic DNA (Figs. 3A and 2, LMP7
BamHI blot). Two LMP7-positive fragments were
identified in cosmid 32, and sequence of the
XbaI-BamHI fragment revealed LMP7 exon
4 (Fig. 3C, accession AF357928). Exon 5 was identified by
Southern blotting and PCR with exon-specific primers and sequencing of
the product (data not shown). An exon 6 (3' UTR)-specific probe did not
hybridize to cosmid 32; furthermore, none of the fragments upstream
from exon 4 hybridized with the full-length LMP7
probe, and since no other non-MHC-linked bands were detected on the
genomic Southern blot with the probe, this gene must be truncated. PCR
was then performed with cosmid 32 as template and primers designed to
amplify the UAA 1 to 3 domain coding regions. The
sequence of this class I gene in cosmid 32 (UAA-NC2) has two
stop codons in the exon encoding the 2 domain and one in the 3
domain-encoding exon (Fig. 4), and the
intron length and sequence between the 1 and 2 domain-encoding
exons are also not conserved compared with UAA,
UAA-NC2 having 326 bp rather than 108 bp in UAA.
In summary, this cosmid clone contains a pseudogene related to class Ia
and an incomplete LMP7 gene, and this syntenic cluster is
not linked to the functional MHC.
|
|
C) and
shown to bear class Ia genes with sequences precisely matching the two
expressed UAA cDNAs expressed by shark Yellow, the presumed
alleles (UAA03 and UAA05, Fig. 4). To identify
the weakly hybridizing LMP7 band in each cosmid, the 1.9-kb
ClaI/SmaI DNA fragment was subcloned, and the
sequence revealed an LMP7-like exon 1 (accession
AF357927). Double digestions of the cosmid clones with BamHI
and EcoRI demonstrated that the LMP7-like genes
on cosmid clones 19 and 65 are the U and
L bands, respectively, detected on the genomic blot in Fig. 2 (Fig. 3B). As expected from mapping of the cosmids (the polymorphic BamHI site downstream of LMP7-like exon 1), shark Yellow, like the mother of the family shown in Fig. 2, has both bands
( U/L, Fig. 3B). The bands present in shark Yellow and in some of the fathers of the family encode at least one complete LMP7-like gene (there are additional bands not shown in Figs. 2 and 3B). As with the LMP7 gene in cosmid 32,
LMP7-like is
a pseudogene, as only exon 1 was identified from sequencing analysis.
In summary, the cosmid clones bearing the nurse shark MHC-linked
classical class I genes were isolated, and such genes are closely
linked to proteasome gene fragments.
To prove conclusively that some animals lacked a complete
LMP7-like gene, we performed Northern blots with PBL RNA
from eight siblings, three positive (groups d and h) and five negative
for the set
of bands (Fig. 2 and Table II), and with shark Yellow, which is also positive (Fig. 3B). Only those animals that were band positive
on the Southern blot expressed LMP7-like transcripts, while
all nine animals expressed LMP7 (Fig. 5A).
|
set
of LMP7-like bands detected by Southern blotting (Figs. 2 and 5B) did not express this gene (Fig. 5B).
However, we were surprised to find that only four of the wild animals
expressed LMP7, correlating with the presence of a 2.8-kb
RFLP detected on the LMP7 Southern blot (arrow in Fig. 5B). Thus, from the family study and examination of the wild
animals, functional LMP7 and LMP7-like behave
like alleles (pseudoalleles), but LMP7 and
LMP7-like gene fragments are nevertheless present in all
haplotypes (Figs. 2, 5B, 9, and 10). The presence of two
LMP7 alleles with quite different sequences is reminiscent
of the situation in Xenopus, in which ancient
LMP7 lineages have been detected (23) (see
below). Other nonclassical class I
In our previous study, we isolated a nonclassical class I gene
(UAA-NC1, accession AF220360) by PCR with the primer set
used to amplify UAA from genomic DNA (1).
UAA-NC1-specific bands segregating in the family were
detected (Fig. 6), but the
UAA-NC1 groups matched neither the MHC groups nor the
non-MHC-linked TAP2 and LMP7 RFLP (Fig. 2 and
Table I). To further understand the UAA-NC1 protein structure, the
nurse shark spleen cDNA library was screened, resulting in isolation of
a 2072-bp clone. Amino acid identity to Gici-UAA is 81% in
leader, 61% in 1, 75% in 2, and 89% in 3 domains (Fig. 4).
There is a 186-bp short interspersed elements retrotransposon
element in the 3' UTR not detected in UAA, and this element
may have been involved in movement of UAA-NC1 out of the MHC
(42). The beginning of the transmembrane domain is highly
conserved, but there is a 49-nt deletion compared with UAA,
resulting in a downstream frameshift (UAA-NC1a, Fig. 7). In addition, a clone that was 3 nt
shorter at the start of exon 6 was isolated from the same cDNA library
(UAA-NC1b). Since the deletion occurred near the exon-intron
boundary, PCR was performed to amplify the region encompassing exons 5
through 6 from genomic DNA. The 1-kb fragment so amplified
revealed that all the missing nucleotides were indeed present in
UAA-NC1 genomic DNA, and the intron sequence was highly
conserved between UAA and UAA-NC1. However, 3 nt
were deleted from the middle of exon 5 (shown as a bracket above the
sequence), and there are some nucleotide replacements (asterisks
above the sequence) that might promote alternative splicing of
UAA-NC1 transcripts (Fig. 7).
|
|
). The 49-bp deletion
resulting in premature termination in UAA-NC1a provides a
transcript encoding the first 10 (of 24) hydrophobic aa in the putative
transmembrane region. Thus, perhaps like proteins translated from
differentially spliced class I transcripts in mammals (43, 44), different forms of the shark protein might be found.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). In groups
d, e, g, h, k, l, and m, the LMP2 U band seems to segregate
with the very top band (15 kb in Fig. 2, e.g., compare siblings 4
and 5 in which this uppermost band is approximately half of the
intensity in sib 5), but the LMP2 L band is found only in
groups e, g, k, l, and m, suggesting the presence of additional genes
or gene fragments. Whether there is a second expressed LMP2
gene in elasmobranchs, and whether this gene is related to the novel
teleost LMP2 (PMSB11) gene (14, 15, 40, 41) requires further study.
|
C), this gene
arose early in evolution and may be found in all elasmobranchs. Despite
its apparent early emergence, a complete LMP7-like gene is
not found in all haplotypes, and the functional gene is likely to be
allelic with LMP7 (Figs. 5, 9, and 10). It was already recognized that
shark LMP7 and LMP7-like proteins differ in residues near their
respective catalytic sites (31), implying that the
spectrum of peptides might be altered in animals that express both
LMP7 genes compared with those with only one. Fig. 11 displays stereoimages of LMP7,
LMP7-like, and the constitutive enzyme LMPX, modeled onto the crystal
structure of yeast LMPX (45). Residues in red include the
active site Thr-1, the obligatory (proton acceptor?)
Lys33, and Met45, which is
believed to coordinate chymotryptic-like cleavages (after hydrophobic
residues) in this active site (46). Five residues in
positions 28 (green), 29 (yellow), 30 (blue), 31 (black), and 32
(purple) (Ser-Gln-Thr-Val-Lys, Ser-Val-Asp-Ala-Asn, and
Thr-Gln-Met-Phe-Lys in LMPX, LMP7, and LMP7-like, respectively) are
diverse among the three proteins (31), and their positions
relative to the active site are displayed. Note especially the
proximity of the bulky Phe31 (black) to
Met45 in LMP7-like as compared with
Ala31 in LMP7 and Val31 in
LMPX. We think that it would be surprising if these differences in the
homologous catalytic subunits did not alter the spectrum of peptides
produced by proteasomes containing them. Amazingly, for species
separated from each other by 500 million years, the two old allelic
lineages of Xenopus LMP7 genes also differ from each other
primarily in the vicinity of the catalytic site, one lineage having
Phe31 and the other Ala31 (Ref.
23 and Y. Ohta, S. J. Powis, and M. F. Flajnik, manuscript
in preparation). Although we are just beginning our analyses of the
inducible proteasome genes in nonmammalian vertebrates, based upon 1)
the very old gene/allelic lineages for several MHC genes (22, 23, 24, 47); 2) different proteasome genes, in particular MHC haplotypes
(23 ; this study); and 3) the close linkage of proteasome
genes to classical class I (6, 11, 13, 15, 17), one can
propose that the generation of diverse sets of peptides by
immunoproteasomes may be more crucial for adaptive responses in these
nonmammalian species than in mice or humans.
|
|
1 helix, one having 77DILSK and the other
77KTALE (bracketed in Fig. 4). This polymorphic
region itself could be recognized by different sets of NK cell
receptors (e.g., see Ref. 48 for NK receptor interactions
with a human killer inhibitory receptor), or perhaps different
arrays of peptides generated, depending upon the constellation of
inherited proteasome genes, may alter the orientation of this class I
helical region recognized by TCR and/or NK cell receptors. It appears
that the p1 and p2 haplotypes were inherited from one father, but the
p2 haplotype was more successful than p1 in being perpetuated; the p2
haplotype in the d and h groups had RFLP for LMP2,
LMP7, and TAP2, and contains a functional
LMP7-like gene. Perhaps such a haplotype provided an
advantage to offspring through expression of the LMP7-like
gene, the particular class I allele, or other genes in the haplotype,
so that survival was enhanced, but there are many other explanations
for this finding that must be explored.
To date, four class I genes have been isolated from the nurse shark,
one classical gene, and three nonclassical class I genes. The
nonclassical UCA (formerly Gici-11
(34)) was the first gene to be cloned, and it is expressed
in a similar fashion to UAA; unfortunately, we have not been
able to map this class I gene because of its low level of polymorphism.
UAA-NC1, isolated by a PCR "accident" with primers
designed for UAA, was shown in this study not to map to MHC,
despite its high sequence similarity to UAA. Some
UAA alleles are more similar to UAA-NC1 than
others (seen as cross-hybridization in groups d and h in Fig. 6),
suggesting that UAA-NC1 arose from one particular
UAA lineage. The third nonclassical class I
UAA-NC2 was found in a cosmid clone bearing a truncated
LMP7 gene. This pseudogene complex was probably the result
of recent duplication event, which indicates that MHC class Ia was
closely linked to LMP7 in the MHC. In fact, all of the class
I genes found on the cosmid clones isolated to date are linked to an
LMP7 (or 7-like) gene fragment. In the teleosts
Fugu (13, 40) and zebrafish (15, 41), class Ia genes are also physically linked to
LMP7 genes (or gene fragments), suggesting that this close
physical linkage is ancient.
Based upon Southern blotting analyses with a variety of probes under low stringency conditions, UAA may be the only class I gene linked to MHC. The same feature is found for the Xenopus MHC, in which only one class I gene of the classical type is MHC linked (49), while a large set of linked nonclassical class I genes is found in another gene complex half a chromosome away (50, 51). In chickens, two class I genes (one highly expressed) sandwich the TAP1 and TAP2 genes, and Kaufman and colleagues (11, 21, 52) have proposed that the tight linkage has allowed the particular class I and TAP alleles to coevolve; furthermore, like in frogs, several nonclassical chicken class I genes are found a great distance away from the true MHC on the same chromosome (53). In the banded houndshark Triakis scyllium, there are one or two classical class I genes depending on the haplotype examined (54), and the majority of teleost species examined rigorously (except cod (29) and cichlid (25, 26)) have relatively few class I genes in the class I region. Previously, others and we have suggested that close linkage of class I/proteasome/TAP genes constitutes a class I region, in which the genes involved in class I processing and presentation coevolve (10, 21). A large number of class I genes are not optimal in this situation, as this would probably result in rapid evolution of class I by gene conversion or other recombinations known to occur in mammals (55). The nurse shark seems to have carried this feature to the extreme, with even recently duplicated class I genes being exiled from the MHC proper. Thus, we predict that the shark class I genes will also be closely linked to and coevolving with TAP and proteasome genes in a true class I region, consistent with our preliminary analysis of the cosmid clones.
Other cosmid clones containing TAP and proteasome genes have been isolated, and we are currently sequencing and analyzing such clones (in collaboration with T. Shiina and H. Inoko, Tokai University, Isehara, Japan). Examination of these clones will be helpful in comparing with other species to infer the primordial MHC organization. For example, recent work has shown that the third inducible proteasome gene MECL-1, while not linked to the MHC in mammals, is found in the teleost MHC class I region (13, 40, 56), consistent with predictions of large scale duplications being involved in the genesis of MHC (57); linkage of this gene, or other proteasome genes, to shark MHC will further test the duplication model. Since shark classical class II genes are linked to class I, it will be interesting as well to contrast the organization of all shark MHC regions with those of teleosts, in which the classical class II genes have been expelled to another region of the genome (14, 18). Finally, we may uncover genes in the shark MHC involved in Ag presentation that are no longer found in MHCs of more recently derived vertebrates.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Martin F. Flajnik, Department of Microbiology and Immunology, University of Maryland School of Medicine, 655 West Baltimore Street, BRB Room 13-009, Baltimore, MD 21201. E-mail address: MFlajnik{at}som.umaryland.edu
3 Abbreviations used in this paper: RE, restriction enzyme; RFLP, restriction fragment-length polymorphism; UTR, untranslated region.
Received for publication May 2, 2001. Accepted for publication November 2, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
(2)-microglobulin sequences in Atlantic cod reveals an unusually high number of expressed class I genes. Immunogenetics 50:49.[Medline]
and LMP2: complementary DNA cloning and linkage analysis with MHC in lower vertebrates. J. Immunol. 159:734.[Abstract]
proteasome subunit cluster in the Japanese pufferfish (Fugu rubripes). J. Immunol. 165:4446./transporter associated with antigen processing-2 gene cluster and identification of five new proteasome subunit genes. J. Immunol. 163:2657.This article has been cited by other articles:
|
|
 |
|
  Y. Ohta, W. Goetz, M. Z. Hossain, M. Nonaka, and M. F. Flajnik Ancestral Organization of the MHC Revealed in the Amphibian Xenopus J. Immunol., March 15, 2006; 176(6): 3674 - 3685. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. A. Moon, S. M. Veniamin, J. A. Parks-Dely, and K. E. Magor The MHC of the Duck (Anas platyrhynchos) Contains Five Differentially Expressed Class I Genes J. Immunol., November 15, 2005; 175(10): 6702 - 6712. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Ohta, E. Landis, T. Boulay, R. B. Phillips, B. Collet, C. J. Secombes, M. F. Flajnik, and J. D. Hansen Homologs of CD83 from Elasmobranch and Teleost Fish J. Immunol., October 1, 2004; 173(7): 4553 - 4560. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Terado, K. Okamura, Y. Ohta, D.-H. Shin, S. L. Smith, K. Hashimoto, T. Takemoto, M. I. Nonaka, H. Kimura, M. F. Flajnik, et al. Molecular Cloning of C4 Gene and Identification of the Class III Complement Region in the Shark MHC J. Immunol., September 1, 2003; 171(5): 2461 - 2466. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. P. Kruiswijk, T. T. Hermsen, A. H. Westphal, H. F. J. Savelkoul, and R. J. M. Stet A Novel Functional Class I Lineage in Zebrafish (Danio rerio), Carp (Cyprinus carpio), and Large Barbus (Barbus intermedius) Showing an Unusual Conservation of the Peptide Binding Domains J. Immunol., August 15, 2002; 169(4): 1936 - 1947. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
DNA used as size standard is noted
on the right. B, Southern blot comparing
genomic DNA and cosmids 19 and 65. A total of 5 µg of mother,
siblings 15 and 37, shark Yellow genomic DNA, and 50 ng of cosmid DNA
was loaded and hybridized with the full-length LMP7-like
probe. BamHI/EcoRI double digests were
used to identify band 
, respectively. The second
LMP2 gene (fragment) is shown as a lightly shaded box.
The order of genes and intergenic distances are not to scale (and are
unknown). TAP2, LMP7, and
LMP2 were definitively shown to map to MHC only in
haplotypes from the P2, P4, and P5 groups.
