| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
* UFR de Biologie, Universite Joseph Fourier-Grenoble 1, France; and
Department of Biochemistry and Molecular Biology, Medical University of South Carolina, Charleston, S.C. 29425, USA.
1Correspondence: Department of Biochemistry and Molecular Biology, Medical University of South Carolina, 171 Ashley Ave., Charleston, S.C. 29425, USA.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
Key Words: relaxin coding region • DNA polymerase • genomic potential hypothesis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
, has recently been shown to also exist in the ovaries of
marsupials (2)
and Chondrichthyes (3)
. In human males, relaxin appears
to be produced in the prostate, where it may function as a sperm
motility factor (4)
. Serological studies suggest that relaxin may exist
in protozoa (5)
. Georges et al. (6
, 7)
reported the first detection of
relaxin-like material in urochordates in 1990. We report the
identification of a relaxin sequence in the genomic DNA and mRNA of the
tunicate Ciona intestinalis and a partial amino acid
sequence of the tunicate relaxin. These results provide compelling
evidence for the structural identity of the tunicate relaxin with the
porcine hormone. The fact that relaxin is expressed in C.
intestinalis only during the spawning period leads to the most
interesting notion that relaxin may be associated with, if not
functional in, the propagation of Urochordata. |
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
Genomic DNA was extracted from one ovary (~200 mg) per experiment with a DNA extraction kit (Stratagene, San Diego, Calif.) and kept in TE (tris or tricine/EDTA) buffer at 4°C. Several extractions were made from gonads harvested during the spawning period from June to July as well as during a quiescent period (March to the end of May).
Total RNA was isolated by a single-step extraction from 8 g
of ascidian ovaries (~40 ovaries) according to Chomczynski and Sacchi
(8)
. The purification of poly(A) + RNA was accomplished by affinity
chromatography on oligo(dt) cellulose, as reported by Aviv and Leder
(9)
. Aliquots of mRNA were kept in 70% ethanol and frozen at -80°C.
Total RNA was isolated from 20 mg of control tissue, using the `RNA
now' (Stratagene kit) or RNeasy mini kit (Qiagen, Chatsworth, Calif.).
Electrophoresis of RNA samples (~5.9 µg) was performed in duplicate
on 0.66 M formaldehyde/agarose gel together with 1.9 µg of control
RNA (Clontech, Palo Alto, Calif.). One part of the gel was stained to
check the purity of the RNA, the other was hybridized.
Amplification procedures and oligonucleotide sequence
Since no sequence was known for the ascidian relaxin, primers
were designed as a consensus of degenerate primers based on lower
vertebrate insulin and relaxin structures. The primers were then
reworked with the PC/gene program according to the sequence of pig
relaxin (10)
and generated by Genosys or Eurogentec programs.
Taq and TaqStart antibody, vent, Takara, or Pfu DNA polymerase were
tested. The cycling parameters used with Pfu DNA polymerase on a
minicycler PTC-150 (MJ Research, Watertown, Mass.) were 94°C, 45 s, 35–40 cycles at 94°C for 45 s, 50°C, 55°C, or 60°C for
45 s, 72°C for 2 min, and then 72°C for 10 min. The amplified
products were analyzed on low melting gel of 1% to 1.8% LMP agarose
(Life Technologies, Inc., Grand Island, N.Y.) in recycled TAE buffer;
the gel was stained with ethidium bromide for 15 min, destained for
2 x 10 min each in distilled water, and observed under UV light
(312 nm). Bands were excised and eluted with Agarose spin columns
(Supelco, Bellefonte, Pa.); the product was reamplified with the same
primers and either sequenced directly (Genome Express) or after
cloning, using the Clone-amp system (Life Technologies) or PCR-Script
Amp SK (+) cloning kit and Screen Test recombinant screening kit
(Stratagene). The 5' RACE and 3' RACE experiments were performed
according to the technique described by Chenchik et al. (11)
with the
Marathon cDNA amplification kit from Clontech.
For hybridizations of Northern or Southern blots, Polarplex
chemiluminescent blotting kit (Millipore, Bedford, Mass.) was used
after overnight transfer in 10x SSC buffer by capillary action onto
Immobilon-S membranes. The membrane was then air-dried and fixed for 4
min under UV light at 254 nm. Hybridization with the porcine
preprorelaxin probe was performed at 55°C or 60°C. The probe,
designed according to porcine preprorelaxin, was a gift from Dr. Kwok
(12)
.
Radioimmunoassay of tunicate relaxin
Specific anti-relaxin antibodies were prepared in rabbits
against a receptor binding site-containing synthetic decapeptide
covalently linked to a linear antigen presenting unit (patent pending).
The antibodies obtained reacted equally well with porcine and human
relaxin but not with the relaxin-like factor, insulin, or bombyxin.
Radioiodine-labeled porcine relaxin tracer was used as the competitor
when testing tunicate extracts for relaxin.
Isolation and partial sequence analysis
Lyophilized gonadal tissue (~100 mg) was suspended in 0.5 M
HCl for 6 h at 4°C. Thereafter, cold acetone was added up to
70% of the total volume. After 12 h, the pellet was collected by
centrifugation and discarded, the supernatant adjusted to pH 7.0, and
desalted on a 2 x 30 cm G25 column. The first peak, which
contained the molecules that reacted with anti-porcine relaxin
antibodies, was collected and lyophilized. Separation on a Sephadex G50
SF column caused the relaxin antigenicity to move to the end of the
elution curve, indicating that its approximate molecular size was
relaxin-like (~6 kDa). Subsequent separation on a quaternary amine
column (0.5 x 10 cm, Pharmacia, Piscataway, N.J.), using a
Pharmacia FPLC, yielded relaxin-like material in the breakthrough peak
whereas separation on a strong cation exchanger moved the ascidian
relaxin to the end of the linear gradient, starting with 50 mM sodium
acetate (A) to acetate 0.5 M in NaCl (B) at a flow rate of 0.5
ml/min.
The analytical high-performance liquid chromatography (HPLC)2 (ABI 130) was chosen as the next `miniature preparatory' step. A linear gradient was programmed to run for 50 min at 100 µl/min, beginning with 0.1% TFA in 25% acetonitrile and ending with 0.1% TFA in 80% acetonitrile. Fractions were collected manually in tubes containing 20 µl of 80% acetonitrile in 0.1% TFA to minimize adsorption of relaxin to the tube wall and assayed for relaxin immune activity with anti-porcine relaxin antibodies; the relaxin-containing fractions were lyophilized. The immune-active material was redissolved in a small volume of 10% acetonitrile in 0.1% TFA; 5 µl was injected into an ABI capillary HPLC blotter system-173 and eluted with a 90 min gradient of 15% to 65% acetonitrile in 0.1% TFA. To determine the exact position of the relaxin immune activity, the column effluent was led directly into a series of radioimmunoassay tubes containing assay buffer. The column effluent capillary was moved manually so as to collect the effluent corresponding to exactly 1% of the gradient in each tube. The fractions thus collected could be related precisely to the absorption measured at 210 nm. Subsequently, the relaxin peak was collected from five such runs directly onto a glass membrane and transferred to an ABI Procise sequencer for microsequence analysis. In all, three filters were used as controls to collect the effluent immediately before and after the immune-active fraction.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
have shown by
immunofluorescent techniques that insulin and immunoglobulin F1, but
not relaxin, exist in the intestines of C. intestinalis.
Nor have degenerate primers used in earlier experiments ever revealed
relaxin-like sequences. In contrast, primers designed according to the
porcine relaxin sequence, as shown in Fig. 1
, led to positive results. In general, the same results were obtained
with genomic DNA or cDNA after cloning or by direct sequencing of the
amplified product. These results were obtained in four or five
experiments before they were considered secure.
|
A unique 315 base pair (bp) fragment was obtained by cloning an
amplified product with a sense primer (Fig. 1)
(#1) directed toward the
C-peptide region and an antisense primer (#2) located at the 3' end in
the untranslated region. Again, the same results were obtained by
direct sequence determination of the amplification product.
A 301 bp fragment obtained after amplification of DNA with primers #3
and #4 was subjected to sequence determination. The tunicate sequence
was identical to porcine relaxin, except for position 92, which showed
serine rather then leucine, as in pig relaxin. Haley et al. (10)
reported the occasional exchange of Leu
Ser in position 92 in porcine
relaxin.
The amplifications with primers #5 and #6 led regularly to two bands, one of 700 bp and one of 139 bp. The small band gave rise to an open reading frame for porcine relaxin, whereas the larger band did not. Primers #7 and #8 also led to a 700 bp fragment with no relaxin affinity and to the 193 bp fragment, which again coded for relaxin sequences. At this point, we discovered that in one of ten experiments, glutamic acid B13 was replaced by lysine. This observation suggests that C. intestinales may have two relaxin genes, like humans and mice.
The 5' end was more difficult to obtain. The strategy was to use
the 5' RACE protocol from Life Technologies, starting with a specific
antisense primer #6, followed by a universal primer, and then an
internal antisense primer #8. Two bands (260 bp and 180 bp) were
obtained and each reamplified with two other internal primers #9
(sense) and #10 (antisense). Only one fragment gave an open reading
frame sequence (176 bp) that was related to the porcine relaxin with a
difference of one amino acid, Ile-17 instead of Leu, in the signal
peptide. Position 3 in the signal peptide showed a PheIle exchange
once in ten times. The total sequence of the tunicate relaxin coding
region is shown in Fig. 2
.
|
The mRNA extracted from ovaries of ascidians collected from March to
June in the English Channel did not contain any sequences related to
relaxin, which correlates well with the absence of mature oocytes at
that time of the year. The positive results (at the 5' end) were
obtained only with animals collected in July, when the ovaries had
matured and the animals were ready to spawn. Immunofluorescence studies
have shown that only mature oocytes with well-developed follicle cells
reacted with the porcine anti-relaxin antibody (6
, 7)
.
The sequences of tunicate relaxin coding regions show clearly that there is no intron. This is different from the porcine gene, but identical to the bombyxin gene, which codes for a relaxin-like molecule in the silk moth, Bombyx mori.
Northern analysis with the preprorelaxin probe (not shown) revealed several bands larger than 1 kb. This had already been observed by one of us (D.G.) during a set of experiments with a different species from Hawaii (Herdmania momus). In that case, poly(A) + RNA failed to hybridize with the porcine C-peptide probes, suggesting that H. momus relaxin is different.
Relaxin shows high variability even among closely related mammals. The
startling exceptions have been the relaxins of the whales,
Balaenopterae acutorostrata and
Balaenopterae edeni, which differ from each other
by three residues, whereas B. edeni relaxin
differed at only one position from that of pigs (14)
. Although this
observation has been acknowledged as one of nature's quirks, the
ascidian relaxin indicates that there may be a serious problem with the
concept of molecular genealogy.
The high homology of sequences of pig and ascidian relaxins alerted us to the dangers of contamination. Extreme precautions were therefore taken during experiments. Parallel experiments with intestinal tissues were consistently negative. Moreover, every time two bands were obtained by amplifying cDNA, only one gave a sequence related to relaxin.
The danger of contamination is moderate in DNA work, but it is a serious problem during small-scale protein purification. Consequently, only new column material and glassware were used for the isolation and purification procedures.
The acid-acetone extract of tunicate gonads reacted with anti-porcine
relaxin antibodies, and a dilution curve showed identical slopes
(Fig. 3
), suggesting similarity with porcine relaxin. In fact, throughout all
of the subsequent purification procedures by ion-exchange
chromatography (not shown), the similarity to porcine relaxin became
more obvious. Finally, separation on a capillary HPLC system gave rise
to the data in Fig. 4
, which shows that the immune activity from C. intestinalis
coincides with the elution of a porcine relaxin standard.
|
|
Separation by capillary HPLC yielded a compound peak, which could
be collected in three separate fractions directly onto a glass filter
for sequencing. All three fractions were analyzed, but only the one
that had previously been shown to contain the immune-active component
revealed the sequence of the ascidian relaxin A-chain, which was
identical to the corresponding chain in porcine or whale relaxin
(Table 1
). The B-chain has a glutamine residue at the NH2
terminus, which converts quantitatively to pyrrolidonecarboxylic acid
and is therefore silent during sequencing. Together with the data
obtained by gene analysis and cDNA sequence determination, these
results leave no doubt that the relaxin of the tunicate C.
intestinalis is identical to the circulating form (B29) of porcine
relaxin (Fig. 5
).
|
|
It should be noted that our ability to extract relaxin, however little, and obtain a partial sequence analysis means that the molecule is expressed in tunicates during the reproductive cycle. In the unlikely event that artifacts always made their appearance in the spawning season during several years of cloning and isolation in the laboratory in France, it would be taking caution too far to assume that such a contaminant could also show up in the laboratory in the U.S., this time in its translated form as a protein. Our finding is taking a big bite out of an endearing dogma, and it would have been good to see the whole sequence of the tunicate relaxin; but even with these investigators' considerable experience, it would take a great many animals to improve our data beyond this point.
The phylogenetic distance between tunicates and pigs is simply the whole evolutionary time allotted by the Darwinian model for the evolution of all of the diversity observed today. In contrast, this report not only suggests that the pig and tunicate molecule are the same, but invites speculation that the same hormone might be involved in the reproductive process in two phyla that have been separate for 500 million years. The fact that the tunicate molecule is identical to the most potent of relaxins (in the mouse bioassay) should give pause for rethinking the concept of functional adaptation. It is even more amazing that among land mammals, relaxins vary by ~55% even though their insulin and cytochromes show far less structural diversity.
One must conclude that there are identical relaxins in members of
two different phyla: Vertebrata and Urochordata. Molecular genealogy
based on relaxins would place pigs much closer to the tunicate and
whale than to their fellow land vertebrates. Such a distribution of
homologous proteins does not support a linear succession model
(neo-Darwinian), but favors a clonal model of evolution such as the
genomic potential hypothesis (15)
, which separates to a significant
extent the development of species from that of protein structures.
The genomic potential hypothesis is a polyphyletic clonal model that presumes proteins are derived from normal distributions of abiotically polymerized nucleic acids, and that during the process of cell formation, messages for proteins are incorporated into nascent cells according to proximity to the large pools of abiotically produced redundant nucleic acids. Neighboring cells can end up with a relaxin from the same distribution of sequences, whereas insulin genes may be from different regions. Species, of course, will have the same proteins; beyond that, similarity has no meaning as regards phylogeny.
The absence of an intron has been considered a strong indicator of
polymerase chain reaction artifacts. One must consider that the absence
of an intron in the cDNA may also mean that there is no intron in the
gene, as in the case of the relaxin-like insect hormone, bombyxin II
(16)
.
The identity of whale relaxin with porcine relaxin (14)
has been
regarded as a quirk of nature (17)
and an isolated case, but now
C. intestinalis reopens the problem with even greater
urgency. The implications of our findings have caused us to proceed
with extraordinary care: to the best of our knowledge, porcine relaxin
is indeed tunicate relaxin.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication November 3, 1998.
Revision received December 12, 1998.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Hafner and G. Korthof Does a "500 million-year-old hormone" disprove Darwin? FASEB J, July 1, 2006; 20(9): 1290 - 1292. [Full Text] [PDF]
|
|
|
|
 |
|
 T. Klonisch, C. Froehlich, F. Tetens, B. Fischer, and S. Hombach-Klonisch Molecular Remodeling of Members of the Relaxin Family During Primate Evolution Mol. Biol. Evol., March 1, 2001; 18(3): 393 - 403. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
