Gene 295 (2002) 265–277

www.elsevier.com/locate/gene

Transgenic tilapia and the tilapia genome

N. Maclean*, M.A. Rahman, F. Sohm, G. Hwang, A. Iyengar, H. Ayad, A. Smith, H. Farahmand
Division of Cell Science, School of Biological Sciences, University of Southampton, Bassett Crescent East, Southampton SO16 7PX, UK
Received 10 August 2001; accepted 18 May 2002

Abstract
The tilapia fish (Oreochromis niloticus) has an important place in the aquaculture of the developing world. It is also a very useful
laboratory animal, and readily lends itself to the transgenic technology. Through the use of reporter genes, a range of potential gene
promoters have been tested in tilapia, both through transient and stable expression of the reporter construct. Using the transgenic technology,
growth enhanced lines of tilapia have been produced. These fish have no abnormalities and offer a considerable growth advantage for future
exploitation. It is however crucial that transgenic fish, to be exploited in aquaculture, be sterile, and various methods of achieving sterility are
considered. These include triploidy, gene knock out of crucial hormone encoding genes via homologous recombination, and knock down of
the function of the same genes via ribozyme or antisense technologies.
Transgenic tilapia also offer the potential for exploitation as biofactories in the production of valuable pharmaceutical products, and this is
also discussed. q 2002 Elsevier Science B.V. All rights reserved.
Keywords: Transgenic fish; Reporter gene; Antisense; Ribozyme; Biofactory

1. Introduction tively robust and disease free, and are capable of rapid
growth, being sexually mature at 6/7 months from hatch
Tilapia fish are of huge importance in world fisheries, and marketable at this age. The male fish grows faster and
being the second most important group of food fishes in becomes larger than the female and is the more desirable
the world, next to the carps. There are about 100 species fish at market.
and subspecies, of which Oreochromis niloticus, Oreochro- O. niloticus is also an excellent laboratory animal and
mis mossambicus and Oreochromis aureus, and hybrids deserves to be better known as such. It is a mouth-brooder,
between them, are much the most important. O. niloticus and in aquarium conditions is reproductive throughout the
accounted for a harvest of nearly 0.9 million tonnes in 1998 year. Eggs and milt can be stripped from ripe fish, fertilised
(FAO, 1999), second only to carp as a warm water food fish by hand, and the eggs reared to hatch in small Zugler
and exceeding the harvest of Atlantic salmon, Salmo salar, vessels. Perhaps the greatest drawbacks of the fish as labora-
although the value of the Atlantic salmon catch is more than tory model animals is their aggressivity – individual brood
twice that of the tilapia catch. stock are best kept in isolation – and in genetic terms,
Although native to Africa, tilapias are cultured in much of knowledge of the tilapia genome is rather preliminary, and
Asia and the far East, and really occupy two rather separate far behind Fugu and the zebra fish. Table 1 lists the parti-
market niches, being a poor man’s food fish in much of Asia cular benefits and limitations of tilapia as a model species. It
and Africa, but being a high value food fish in countries such must also be emphasised that fish are ancient and very
as Israel and the Southern United States. Tilapias are omni- diverse in evolutionary terms, so the molecular sequence
vorous, but will subsist well on a vegetarian diet, are rela- homology between different teleost families is often less
than between other vertebrate classes.

Abbreviations: LUC, firefly luciferase; GFP, green fluorescent protein;

NEO, neomycin resistance; G0, first generation transgenic fish; G1, second
generation transgenic fish; MUG, methyl umbelliferone galactoside; GH, 2. Materials and methods
growth hormone; GnRH, gonadotropin releasing hormone; LH, luteinising
hormone; PCR, polymerase chain reaction
Since this paper is in many respects a review of work,
* Corresponding author. Tel.: 144-23-8059-4403; fax: 144-23-8059-
4269. much of which is already in publication, this section will be
E-mail address: [email protected] (N. Maclean). brief.
0378-1119/02/$ - see front matter q 2002 Elsevier Science B.V. All rights reserved.
PII: S 0378-111 9(02)00735-7
266 N. Maclean et al. / Gene 295 (2002) 265–277

Table 1
Tilapia (O. niloticus) characteristics as laboratory model species

1. Warm water required, since fish will not survive much below 118C, and 28–308C is optimal (Ross, 2000)
2. Fresh water required, although O. niloticus will survive in mildly brackish water. O. mossambicus is more tolerant of high salinity than is O. niloticus
3. Reproductive throughout the year, males being constantly capable of being stripped, and females spawning approximately every 3–4 weeks. Eggs can
be stripped from females only minutes before natural ovulation, so it is necessary to recognise when a particular female is ripe by appearance of genital
papilla and ‘nest’ building behaviour with gravel
4. Eggs, available as a few hundred per ovulation, can be reared after fertilisation in small Zugler vessels in which the eggs are constantly agitated by a
water current. Hatching occurs in 72–96 h, and embryos are nutritionally dependent on the yolk sac until approx. day 10
5. Very robust and disease free
6. Adults remain reproductive for about 2/3 years
7. Individuals are very territorial and aggressive and so brood stock fish must be segregated
8. Adult fish are large, up to 25 cm in length at 8 months, and so a large aquarium system is needed

2.1. Transgenic fish used 2.3. Detection of lacZ by methyl umbelliferone galactoside
(MUG) assay
The transgenic fish described were produced by micro-
injection of fertilised eggs as discussed in Rahman and The MUG assay was carried out by the method described
Maclean (1992). Following injection of linear transgene on Hoefer’s Technical Bulletin #129 by using a Hoefer TKO
molecules into the egg perinuclear cytoplasm, eggs and 100 minifluorometer. In the MUG assay, 4-methyl umbelli-
embryos were kept in mini-Zugler vessels until yolk sac feryl-b-d-galactoside was used as a substrate which was
absorption, with a constant upward current of water to then cleaved by b-galactosidase to yield the fluorescent
keep the eggs and embryos agitated. Positive transiently molecule 4-methylumbelliferone (7-hydroxy-4-methylcou-
expressing fish were detected by assays for the reporter marin, MU). After excitation at 365 nm, 4-methylumbelli-
gene expression, while fish with stable integrated transgene ferone emits light at a wavelength of 460 nm. This assay is
copies were detected by initial polymerase chain reaction sufficiently sensitive to detect picogram quantities of b-
(PCR) of fin clips and subsequent Southern blots. galactosidase.

3. Results and discussion

2.2. Growth enhanced tilapia
3.1. Transgenesis in tilapia
The growth enhanced tilapia tested in Szarvas, Hungary
carried a single copy of a chinook salmon (Oncorhynchus Transgenic fish are useful for a number of different
tschawytscha Walbaum) growth hormone gene spliced to an reasons. They can provide good models for studies on
ocean pout (Macrozoarces americanus Black and Schnei- gene regulation and gene expression during development,
der) antifreeze promoter co-ligated with a carp beta actin/ transgenesis provides a way of improving fish for aquacul-
beta galactosidase (lacZ) reporter gene construct, integrated ture by conferring desirable genetic traits, and transgenic
into the tilapia genome. The salmon growth hormone gene fish can be exploited as biofactories to produce valuable
construct was kindly given to us by Prof. Choy Hew. Details pharmaceuticals.
of the growth trials of these fish will be found in Rahman et Fish of many species have been used in transgenesis, for
al. (2001). example common carp (Cyprinus carpio) by Zhang et al.

Table 2
Benefits and limitations of tilapia fish as a model organism for the transgenic technology

Benefits
1. Short generation time of 6–7 months
2. Eggs constantly available in a laboratory system
3. Embryos and fry easily reared
4. Embryos semi-transparent, so some reporter genes can be detected without sectioning or sacrifice
5. Eggs fairly easily injected via micropyle, i.e. easier than eggs of salmonids, although the egg is opaque and the chorion tough. However,
injection needles must be mounted in a micromanipulator and each egg individually positioned for optimal needle entry
6. The increased size over medaka and zebrafish means that blood and tissue volumes and sizes are much increased

Limitations
1. Embryos less transparent than those of zebrafish, and so transient expression less easily detected with reporter genes
2. Knowledge of genome and early development is very limited as compared to zebrafish
3. Eggs less easily injected than those of zebrafish, medaka and carp
 N. Maclean et al. / Gene 295 (2002) 265–277 267

(1990), channel catfish (Ictalurus punctatus) by Dunham et those of zebra fish, and the depth of knowledge about the
al. (1987), African catfish (Clarias gariepinus) by Volckaert genome and aspects of molecular development are much
et al. (1994), Northern pike (Esox lucius) by Guise et al. less than in that species. Although egg injection is not as
(1992), rainbow trout (Oncorhynchus mykiss) by Guyomard easy as with carp or zebrafish, it is much easier than with
et al. (1989), Atlantic salmon (S. salar) by Shears et al. salmonids.
(1991), mud loach (Misgurnus mizolepin) by Nam et al. Fish have their own characteristics as a transgenic
(2000), and smaller model species such as goldfish (Caras- system, and these hold true for tilapia transgenesis. Table
sius auratus) by Yoon et al. (1990), and zebra fish (Danio 3 outlines the major features.
rerio) by Stuart et al. (1990).
Transgenic tilapia have been produced by Brem et al. 3.2. Use of reporter genes in transgenic tilapia
(1988) and Martinez et al. (1996) and by Rahman and
Maclean (1992), but the Havana, Cuba laboratory and the Reporters such as chloramphenicol acetyl transferase
Southampton, England laboratory are now the chief centres (CAT), lacZ, firefly luciferase (LUC), and green fluorescent
working on transgenesis in tilapia. Tilapia have much to protein (GFP) have been widely used in fish (see Section 3
offer as a transgenic system and the pros and cons are listed in Iyengar et al., 1996; Maclean, 1998). The neomycin resis-
in Table 2. Particularly noteworthy are the short generation tance reporter/selection system is less useful since the
time, constant egg availability, and comparative ease of egg mosaic distribution and expression in the first generation
injection and rearing of the embryos. Thus the generation transgenic fish (G0) embryos and fish means that the
time is about 6 months, little longer than zebra fish, and the embryos or fry are killed by the neomycin or G418 since
rearing of the fry is substantially easier than with zebra fish. non-transgene-expressing cells are susceptible. Good quan-
However, the embryos are somewhat less transparent than titative assays for GFP expression are difficult to set up,

Table 3
Transgenesis in tilapia fish (many aspects cited hold true for all transgenic fish systems)

1. Although many alternatives to needle microinjection of the fertilised egg have been tried, such as electroporation, gene guns, oocyte injection and
liposome mediated gene transfer via sperm, none have, to date, proved reliably superior to microinjection
2. The fertilisation pronuclei cannot be visualised, even microscopically, so DNA copies have to be injected into the perinuclear cytoplasm of the fertilised
egg. Subsequent uptake into cell nuclei and chromosomal integration is therefore not guaranteed
3. Many transgene copies are required, 5 £ 10 5 in tilapia. This is a balance between injecting too much DNA, which is lethal to egg and embryo survival,
and injecting a low copy number and finding that gene integration is exceedingly infrequent
4. Following embryonic development, a majority of embryos will display transient expression of some transgene copies. This is from transcription and
translation of unintegrated transgene copies. Although the transient expression is mosaic (see 5 below) it is tissue specific and dependent on the
characteristics of the promoter driving the construct. However, transient expression levels are highly variable from embryo to embryo, and so large
numbers of embryos must be screened if data is to be quantitative
5. Expression in the first generation fish grown from injected eggs is invariably mosaic, presumably because the transgene copies are distributed in a
somewhat random way following early cell division in the embryo. Stable expression (see 6 below) in first generation fish is, like transient expression,
invariably mosaic
6. Only a small proportion (1–5%) of injected eggs will produce fish showing stable transgene expression from integrated copies
7. Since even fish with integrated transgene copies have a mosaic pattern of integration, only a subset of G0 fish will demonstrate germ line transmission
8. Once second generation transgenic fish (G1) progeny are produced, the problem with mosaicism is solved, but the proportion of G1 positive transgenics
may be low
9. PCR from fin clip material can be used to identify presumptive transgenic positives, but some will be missed because of mosaicism. PCR should be
followed up with Southern blotting to confirm the presence of the transgene. Some long term persistence of non-integrated copies is possible, so
Southern blotting and germ line transmission are needed to confirm chromosomal integration
10. All transgenic fish are initially hemizygous, that is that integration occurs on only one of a pair of homologous chromosomes. However homozygous fish
can be produced by crossing two hemizygous fish of the same line and identifying the 25% homozygous progeny
11. Transgenes readily form concatamers in fish, and so most transgenic fish will be found to have multiple copy integration events. Multiple integrations
into separate chromosomes are also not uncommon
12. As will be discussed in a later section, there is no equivalent to the mouse embryonic stem (ES) cell system in fish, so gene knock-out and homologous
recombination is presently difficult
13. A range of reporter genes such as CAT lacZ LUC and GFP are useful in tilapia, and some, such as GFP, can be detected in living fish, although
expression is hard to quantify precisely
14. A trained worker producing transgenic tilapia should be able to reach the following targets
(a) Eggs injected per day 250
(b) Eggs hatching to yield viable embryos after egg microinjection 260 out of 100
(c) Eggs developing to show transient expression between 2 and 10 days of embryonic development 240 out of 100
(d) Eggs developing into embryos and yielding fish showing mosaic stable expression at 1 month, 21–5 out of 100
(e) Fish grown from injected eggs which demonstrate some germ line transmission 20.1–1 out of 100 eggs injected
(f) Percentage of positive transgenic progeny from a cross between a stable GO transgenic and wild type fish 20.5–10 out of every 100
(g) Percentage of positive transgenic progeny from a cross between a stable G1 transgenic and a wild type fish 250 out of 100 (i.e. hemizygous)
268 N. Maclean et al. / Gene 295 (2002) 265–277

although it has the merit of being detectable in living 3.2.2. Use of reporter gene constructs to demonstrate
embryos. LacZ has much to commend it and the MUG transgene homozygous integration status
assay provides a sensitive way to measure expression levels, It is sometimes desirable to obtain fish in which both
but unfortunately there is evidence that in some cells and copies of a homologous pair of chromosomes carry inte-
tissues, in fish, as in mice, (Montoliu et al., 2000) the repor- grated transgenes. In a recent paper (Rahman et al., 2000),
ter is capable of silencing the promoter in use and so some we have demonstrated the use of the lacZ reporter to indi-
degree of transgene expression mosaicism results. (Rahman cate the production of homozygous transgenics by crossing
et al., 2000).
The following experiments have been conducted in tila-
pia using reporter gene transgene constructs.

3.2.1. Demonstration of copy number related expression

levels in different transgenic lines
As demonstrated in Rahman et al. (1997) three different
transgenic lines were compared, in which the concata-
merised copy numbers were 19, 34 and 330 copies in each
of the C86, C118 and C58 lines respectively. Mature fish
from each line were dissected and the brains removed for
subsequent lacZ expression level assay. As can be seen in
Fig. 1, the expression levels are broadly in line with the copy
number. The copy number was determined by densitometric
determination following Southern blotting on gels. A more
detailed quantitative assay was carried out by MUG assay of
homogenised G2 and G3 non-mosaic embryos of each line.
In each case, when the expression level in the homogenates
(expressed as 10 3 £ MU units/mg protein) were divided by
the determined copy number, figures of 0.29, 0.32 and 0.23
were obtained for the respective lines C86, C118 and C58,
suggesting good correspondence. This evidence is cited not
to suggest that such a correlation will always obtain in fish –
almost certainly it will not – but to emphasise the usefulness
of the reporter gene assay.
Use of reporter gene constructs to optimise promoter
selection for subsequent inclusion in new gene constructs
or to demonstrate expected patterns of tissue specific
expression.
When new promoters are chosen to drive particular trans-
gene sequences, it is crucial to determine what length of
upstream regulatory sequence will give maximal expres-
sion. This is usually a matter of including any available
enhancers and avoiding any silencer elements. Such assays
may of course be readily carried out in tissue culture cells
where the promoters are ubiquitously expressed, but careful
analysis of transient or stable expression in fish is often
necessary when no relevant tissue cell line is available.
Thus, we have recently helped to characterise a tilapia
L18 ribosomal protein gene by testing promoter regions of
differing lengths in CV1 cells (monkey kidney cell line), and
by measuring transient expression in both tilapia and zebra- Fig. 1. (A–C) Brains of 7 month old G2 fish of the C86, C118 and C58 lines,
fish embryos (Molina et al., 2001), in combination with a respectively, following dissection and staining after removal with X-gal
lacZ reporter gene. A tilapia heat shock HSP70 gene has (magnification £ 10). Note the strong expression in one area of the tele-
ncephalon (left of the picture) in all three lines, contrasting with the limited
also been isolated and characterised, (Molina et al., 2000),
expression in the cerebellum and optic tectum evident only in the C58 line
although in this case the expression analysis was carried out (C). Copy numbers of the reporter gene concatamer in the three lines are 19,
with transient expression in zebrafish embryos, rather than 34 and 330, respectively. (Reproduced with permission from Transgenic
with tilapia embryos. Research).
 N. Maclean et al. / Gene 295 (2002) 265–277 269

together two hemizygous transgenics of the same line. The have not been published for these fish. There is evidence from
authentic transgene homozygosity was subsequently proven recent reports that some of these lines of growth enhanced
by showing that, when crossed with a wild type tilapia, tilapia have been commercially exploited and made available
100% of hemizygous transgenic progeny were obtained. for food in local markets (de la Fuente et al., 1999).
The results of the growth enhancement work done by the
3.3. Use of transgenic technology to produce improved lines Southampton group have already been published (Rahman
of tilapia for exploitation in aquaculture et al., 1998, 2001; Rahman and Maclean, 1999), at least as
far as non-commercial trials are concerned, and the data will
It has been clear for some time that the transgenic tech- simply be summarised here. It is important to stress that a
nology could be used to produce fish with improved genetic number of different lines of fish were produced, and all
traits such as faster growth, improved cold tolerance or resulted from coinjection of the GH construct along with
freeze resistance, increased disease resistance, or enhanced another sequence, carp beta actin promoter driving a lacZ
tolerance of salinity. In this paper the production of growth reporter (Rahman et al., 1997). This coinjection strategy
enhanced lines of tilapia will be used to illustrate this aspect. was devised in order to increase the integration frequency
of the GH transgene, and, insofar as this experiment was
3.3.1. Production of growth enhanced lines of tilapia concerned, it succeeded. Although three transgenic lines
Ever since Palmiter et al. (1982) published their results were produced with integrated copies of the chosen GH
showing dramatic growth enhancement in transgenic mice, construct, only one, the C86 line, has been extensively
fisheries biologists have known that a parallel success was tested. This line contains a single integrated copy of the
possible in fish. However, the most impressive results in GH construct (Rahman and Maclean, 1999) and was chosen
transgenic fish were not published until 1994, when Devlin partly because the level of heterologous GH produced was
et al. (1994) published the results of their experiments with low (1.02 ng/ml of serum), and it was assumed that better all
Pacific salmon. In salmonids, using heterologous growth round performance would be attained with low level
hormone genes spliced to strong promoters of fish origin, increase of GH than with dramatic increase. Certainly in
it proved possible to show more than 10-fold growth large mammals dramatic over-expression of GH can lead
enhancement even with G0 fish, as compared to non-trans- to serious abnormality (Cameron et al., 1994).
genic siblings. Following laboratory-based trials with the C86 line
Since production of growth-enhanced lines of fish is (Rahman and Maclean 1999), more extensive trials were
clearly directed at the goal of improving species for use in carried out at a Fisheries Station at Szarvas, Hungary, in
aquaculture, it is obvious that gene constructs should be the winter of 1997/1998. The choice of this site resulted
designed with this in mind, avoiding as much as possible partly from the availability of local expertise in running
the use of sequences of viral origin and using whenever fish growth trials and partly from effective physical contain-
possible sequences from the same or related fish species. ment. The latter resulted from the availability of warm
The design of growth enhancing gene constructs has geothermal water (which was used via a heat exchanger to
followed to a great extent the experimental plan originally warm good quality water), surrounded by cold water in local
pioneered by Palmiter et al. (1982), namely to use a growth lakes and rivers. This combination ensured that, in the unli-
hormone coding sequence from a species closely related to kely event of fish escaping, they would not survive in the
the target animal, and to drive this with a promoter which locally cold water which surrounded the site. The results of
will ensure synthesis in a large organ such as liver, or with a the trials are reported in Rahman et al. (2001) together with
ubiquitous promoter ensuring widespread growth hormone separate nutritional trials carried out on the same line of fish
(GH) synthesis. at the Institute of Aquaculture, Stirling, Scotland.
In tilapia the work has been concentrated in two labora- In summary, the Szarvas trial lasted for 7 months and
tories, that of de la Fuente in Havana, Cuba (Martinez et al., involved approximately 400 transgenic and 400 non-trans-
1996) and our own laboratory in Southampton, England. genic fish, reared both separately and together. The overall
The Cuban work has used a tilapia GH complementary difference in mean mass was 2.5 times (see Fig. 2), the
DNA (cDNA) driven by a human cytomegalovirus (CMV) transgenic tilapia having a mean mass of 653 g as compared
promoter, whilst the Southampton work has used a gene to 260 g for the non-transgenic fish. No serious abnormal-
construct originally made by Prof. Choy Hew in Toronto ities were evident, and 80 transgenic and 80 non-transgenic
(now in Singapore) and kindly gifted to us. This consists of a fish were dissected at the end of the period to determine
GH cDNA from chinook salmon, driven by a promoter from whether internal abnormalities might have arisen. None
an ocean pout antifreeze gene. (See Du et al., 1992 for such were evident.
details of the original construct and Rahman et al., 1998 Other parameters of interest are that food conversion effi-
for details of the use of the construct in tilapia). ciency was 20% better in the transgenics, and a digestibility
The growth enhancement obtained in tilapia by Martinez trial suggested that the transgenic tilapia were more efficient
et al. (1996) is approximately 2-fold (1.8 times control) when utilisers of protein, dry matter and energy. They also had a
fish are 7 months old, but data from extensive growth trials higher moisture content and lower ash and crude protein
270 N. Maclean et al. / Gene 295 (2002) 265–277

Fig. 2. Photograph showing three 30 week old transgenic fish (left) and three 30 week old non-transgenic full sibling fish (right) from the C86 line. The average
mass of the transgenic and non-transgenic fish in this group was 653 and 260 g, respectively. The fish illustrated were representative of the entire batch.
(Reproduced with permission from Fish and Fisheries).

content and crude lipid content in the flesh than non-trans- recorded improvements obtained by non-transgenic genetic
genics on the basis of a small number of fish analysed. selection are of the order of 1.5–2.0 times (ICLARM Report
Rather similar data to the above have been obtained in http://www.egiar.org/iclarm/gclams/reshigh.htm) the yield
studies on growth enhanced transgenic salmon. as compared to local strains.
From the above evidence it may be concluded that trans- Our laboratory is now engaged in trying to produce new
genic growth enhancement does produce tilapia with desir- strains of transgenically growth enhanced tilapia without
able faster growing properties. The growth rates measured using reporter gene coinjection and use of ‘all-tilapia
are almost certainly not maximal, since only one transgenic constructs’ with tilapia promoters driving a tilapia GH
strain was studied in depth, and there is evidence to suggest gene (see Fig. 3). Such new strains would therefore become
that the transgenic fish were slightly underfed towards the recognised as autotransgenics rather than allotransgenics in
end of the trial period. Of course tilapia growth rate can also the terminology suggested by Beardmore (1997).
be improved by conventional selection, but the best It is obvious that no GM (genetically modified) fish can

Fig. 3. Diagram of an ‘all-tilapia’ GH construct in which a tilapia GH coding sequence is driven by a tilapia L18 (ribosomal protein) promoter. Black boxes
represent the exons. Thin lines represent the introns. Diagram is not to scale.
 N. Maclean et al. / Gene 295 (2002) 265–277 271

be safely exploited in aquaculture unless complete physical banishes fertility. If triploidy were to be used in this way
containment were available in an isolated land-locked lake, to reduce the fertility of transgenically growth enhanced
or if the fish could be biologically contained as a result of tilapia, then a further problem is encountered, since triploid
watertight sterility. This will be discussed in the following transgenics may be actually less growth-enhanced than
sections. diploid transgenics (Razak et al., 1999). In these growth
trials it was found that at 8 months of age normal non-
3.4. The need for sterility transgenic tilapia averaged 39.9 ^ 5.6 g, transgenic diploids
of the C86 line averaged 174.9 ^ 17.6 g while transgenic
As emphasised in the preceding section, sterility is a triploids of this line averaged 138.1 ^ 9.7 g. However, it
necessary adjunct to the exploitation of transgenic tilapia should be stressed that in this study all the triploids were
unless completely secure land locked facilities are available. from fertilised eggs resulting from sperm from a transgenic
In addition, sterility is an important parameter in its own male being used to fertilise the eggs of a wild type female.
right, even aside from its use for containment of GM fish. The triploids therefore carried only one out of three chro-
Tilapia itself is often a pest fish when released into large mosomes with an extra growth hormone gene. If the cross
mixed fresh water fishery waters, since its tendency to had been made the other way, which we propose to try in
precocious sexual maturation means that often large popu- future, the loss of growth enhancement might be much
lations of small sized fish result, and indeed it may be a reduced or absent altogether.
desirable warm water aquaculture species in situations Hamid Farahmand in our group has undertaken the
where its establishment in the wild is undesirable. The production of tetraploid tilapia as a route to the production
problems posed by the escape of Atlantic salmon from sea of 100% triploids following crossing of the tetraploid with
cages and their interbreeding with wild populations in adja- wild type. Evidence for tetraploidy induction has been
cent rivers are well known examples of problems with a obtained using both erythrocyte nuclear diameter and tetra-
different farmed fish species which could at least be partially ploid chromosome spreads (see Fig. 4). However, produc-
reduced by the use of sterile fish. tion of tetraploid tilapia has proved difficult and seems to us
Sterility in fish can be achieved by two different routes, on present evidence to have little commercial future.
namely ploidy manipulation to produce sterile triploids, or
the use of transgenesis to achieve gene ‘knock-out’ or gene 3.4.2. Transgenic reversible sterility
‘knock-down’. Our laboratory is involved with all three To an increasing extent it is becoming clear that, unless
approaches and these will be considered in turn. total physical containment is available, the benefits of trans-
genic manipulation in fish will only be realised if an entirely
3.4.1. Triploidy and tetraploidy in tilapia ‘watertight’ sterility can be induced and used together with
Triploidy provides a way to render fish sterile but unfor- the transgenic improvement, Maclean and Laight (2000).
tunately triploid sterility is rarely 100% (Hussain et al., Such sterility would of course have to be reversible, other-
1991). The lack of complete sterility is usually a result of wise the sterility would involve an equivalent to ‘killing the
some normal diploids hatching from treated eggs and some goose that lays the golden eggs’. Can transgenesis itself be
male triploids having partially functional testes and produ- used to produce reversibly sterile tilapia fish? We believe
cing some viable sperm. This latter problem can be counter- that the answer is affirmative.
acted by using fish which are triploid and all-female through Two different levels of approach are possible with theo-
sex-reversal technology, but unfortunately, in tilapia, the retical reversible sterility induction. One is to ‘knock-out’
male fish are larger and more marketable than the females the genomic copies of a gene whose product is crucial to
and thus all-female triploids are an unattractive commercial sexual development and gonad formation, reversibility
proposition. Triploidy is normally induced in fish through being accomplished by exposure of potential brood stock
temperature or pressure shock to the eggs preventing the to the gene product so that fertility can be restored.
extrusion of the second polar body (Purdom, 1993). As The alternative is to ‘knock down’ the expression of such
mentioned, another problem with such procedures is that a gene by antagonizing the specific message via transgenic
often not all resulting fish are triploid and so even females induction of antisense or ribozyme against the target
can be fertile. An alternative strategy is to produce tetra- message.
ploid fish through shocks at the first mitotic division and Are there potential targets for such transgenic
inducing a state of tetraploidy. Tetraploid fish, unlike approaches? It seems that there are at least two, namely
triploids, are fertile, and produce all-triploid progeny on gonadotropin releasing hormone (GnRH) or luteinizing
crossing with a wild type. Although tetraploid tilapia can hormone (LH). In fish the hormonal cascade which controls
be produced (see later), as discussed above the ‘leakiness’ of gonadal development and gametogenesis involves the
the triploid male makes the technology a less than complete synthesis of GnRH in the brain and delivery of this hormone
answer to the question of sterility in tilapia. to the gonadotropes which synthesise gonadotropin. Fish
Triploidy in tilapia is thus best seen as a partial but not GnRH is a monomer of at least three types in fish. In tilapia
complete answer to sterility. It thus reduces rather than and related fish, the evidence is that the ‘serine 8’ or ‘sea-
272 N. Maclean et al. / Gene 295 (2002) 265–277

Fig. 4. Photograph of mitotic chromosome spread of presumptive tetraploid tilapia embryo following squashing (2n ¼ 44) magnification £ 400.

bream’ type GnRH, synthesised in the hypothalamus, is the 3.4.3. Embryonic stem cells and gene knock out
form which controls gonadotropin release. However, The success of the transgenic mouse system owes much to
another GnRH form, the tryptophan 7 or salmon type the availability of embryonic stem (ES) cells which allow in
appears to have some direct influence also, perhaps partly vitro gene targeting by homologous recombination, followed
as a result of its synthesis in the gonad itself (White and by cell selection and subsequent use of such selected cells to
Fernald, 1998). Thus it is theoretically attractive to attempt produce chimeric embryos and eventual non-chimeric trans-
to knock out or knock down production of one or another or genic lines. Sadly no such ES cell lines are available in tilapia,
both sorts of GnRH, and supply the hormone itself by food or indeed in fish of any species, although in both zebrafish, D.
or injection in order to reverse the sterility. rerio, (Sun et al., 1995) and medaka, Oryzias latipes (Hong et
It is of course gonadotropin itself which controls the al., 1996, 1998) there is experimental promise of such cell
gonad development. This hormone is a polypeptide dimer lines being produced. The lines so far produced in these fish
and exists in two forms, I and II. Type II is responsible for species appear to lack complete totipotency (Hong et al.,
gonad development and this is the form now styled LH. The 1998, 2000) and fail to produce new germ line cells. The
alpha and beta subunits are the products of two different lack of available ES cell lines implies that other ways of
genes and so either gene is a candidate target for knock accomplishing gene knock out must be sought.
out or knock down. In the absence of an effective ES cell system, the only
The conclusion from the above discussion is that both remaining avenue for gene knock out is to achieve gene
salmon type GnRH and sea bream type GnRH make possi- targeting via in vivo homologous recombination, replacing
ble targets in tilapia, but that the ‘sea bream type’ is the the endogenous wild type gene with a mutated silent
more likely to achieve success. In addition, tilapia LH version. However, homologous recombination is a rare
(gonadotropin type II) is a good target in either of its subu- event in transgenic animal systems and requires the use of
nits. We have therefore chosen to target all three tilapia substantial lengths of isogenic sequence identical to the
genes and gene products, that is both of the above GnRH genomic copy to be used to flank the mutated sequence to
genes, and also the tilapia LH beta subunit gene. In all cases be introduced to replace the genomic target sequence.
fertility could be recovered by hormonal administration to Although it is certain that any success achieved would
chosen potential brood stock fish. This follows the model of only involve substitution of one of the two homologous
the hypogonadal mouse (Seeburg et al. 1987), in which a copies, subsequent breeding of homozygous transgenics
mutation in a GnRH gene conferred sterility but adminis- would achieve the desired total gene silencing.
tration of exogenous hormone recovered fertility. Only one system promises to result in efficient homolo-
 N. Maclean et al. / Gene 295 (2002) 265–277 273

gous recombination (HR) in fish, and that is to use the Rad strand or dsRNA (2–4 mg RNA in each case). After 48 h,
51 protein (Sigurdsson et al., 2001), a eukaryotic equivalent muscle blocks of approx. 10 mm 3 were removed from the
of the well known bacterial Rec A protein which has also injected sites and tested for lacZ expression. No obvious
been used effectively to promote HR in both plant and differences in expression level were detected.
mammalian cells. Repeat experiments were carried out in which individual
A combination of a single stranded DNA sequence fish were injected in similar sites on left or right side of the
designed to achieve homologous recombination with, and body with two out of the four compared situations, but again
mutagenic knock-out of, the targeted gene, together with a no differences in lacZ staining were evident.
protein coat of the Rad 51 protein plus replication protein A These experiments with muscle injection were followed
(RPA), holds promise of successful gene knock out in fish. up by egg injection with 10 6 copies of either the sense, anti-
We are therefore in process of injecting tilapia eggs with sense or dsRNA and MUG assays (Rahman et al., 2000) to
such a transgenic complex, designed initially to target the measure levels of lacZ expression carried out on embryo
tilapia serine 8 GnRH gene. homogenates after 24 and 48 h. As seen in Fig. 5, there was
Two probes of 500 and 1000 base pairs have been no difference in the levels of expression (only 50% of
designed. The two probes carry six mutations which intro- embryos are expected to be transgenic since they were
duce five stop codons into the GnRH coding sequence, effec- produced by crossing a hemizygous lacZ expressing fish of
tively destroying the GnRH gene sequence (the wild type the C58 line (with 330 copies of lacZ gene) with a wild type).
sequence of the GnRH extends to ten codons in total). The The proportion of embryos expressing is somewhat variable
last mutation aims to introduce a restriction site for screening but none of the negatively expressing fish injected with
purposes. Before injection into the tilapia eggs following the dsRNA were positive for the transgene on PCR.
methods developed in this laboratory, the probes are made
single stranded by boiling and are then mixed with the recom- 3.4.5. Gene knock down via antisense
binase (kindly provided by Dr P. Sung, Austin, USA) follow- Antisense sequences have now a long history of use in
ing the protocol described by Sigurdsson et al. (2001). After transgenic organisms, especially plants. A useful review of
hatching the eggs are screened by PCR to detect insertion of their use in animal systems is that of Sokol and Murray
the modification into the GnRH locus. (1996). In general it appears that antisense is most effective
against relatively rare messenger RNA species (Branch,
3.4.4. Gene knock down via RNA interference 1998). It is also imperative to use a strong promoter in the
As stated above, the alternative to gene knock-out is gene antisense construct, and of course this must be active in the
knock-down, that is to silence the expression of the target cell type in which the target gene is transcriptionally active.
gene by antagonising the specific messenger RNA. One In our laboratory, antisense constructs have been prepared
possible approach is to use RNA interference or RNAi. against three target genes in tilapia, namely salmon type
This was first reported by Fire et al. (1998) in Caenorhab- GnRH, sea bream type GnRH, and LH beta subunit. All
ditis elegans, where introduction of specific double stranded of these genes are expressed in areas of the brain, and in
RNA was found to inhibit the activity of genes containing addition salmon type GnRH is expressed in gonad (testis)
homologous sequences. This interference is post transcrip- and sea bream type GnRH in spleen (White and Fernald,
tional and appears to be highly effective in C. elegans and 1998). There is some logic in driving an antisense construct
Drosophila (Misquitta and Paterson, 1999) and mice with the promoter sequence which regulates the target gene,
(Wianny and Zernicka-Goetz, 2000). and this we are doing in some cases (see Fig. 6). An alter-
There is a report of specific gene knock down in zebrafish native strategy is to use a strong ubiquitous promoter to
but also of the induction of non-specific defects in zebrafish ensure maximal expression in the target tissue. This strategy
using this approach (Zhao et al., 2001). Since we already is also being followed. The following is a list of the anti-
had lines of transgenic fish possessing and expressing copies sense constructs currently in use in our laboratory.
of a lacZ reporter construct, we have tested RNA interfer-
ence by muscle injection of dsRNA directed against the lacZ † (a) Tilapia beta actin promoter/tilapia LH beta subunit
gene. However, as detailed below, we have not been able to gene antisense.
detect any reduction in the intensity of the lacZ expression † (b) Carp beta actin promoter/tilapia salmon type (trypt. 7)
following this treatment. GnRH gene antisense.
A 3 kb fragment from the bacterial lacZ gene was † (c) Tilapia beta actin promoter/tilapia sea bream type
subjected to PCR and cloned into pGEM-T. Then each (serine 8) GnRH gene antisense.
strand was in-vitro transcribed using SP6 and T7 primers † (d) Tilapia serine 8 promoter/tilapia serine 8 GnRH gene
and the RNA transcript annealed to provide the dsRNA. antisense.
Fish of the C118 line, carrying 34 copies of lacZ in a single † (e) Tilapia histone 3 promoter/tilapia serine 8 GnRH gene
concatamer, were subjected to intra muscular injections of antisense.
the dsRNA at marked positions above the lateral line. Fish
were injected with TE buffer alone, or sense strand, antisense Constructs (d) and (e) are not yet tested.
274 N. Maclean et al. / Gene 295 (2002) 265–277

Fig. 5. Effect of RNA interference against lacZ. Histogram in which each bar represents the level of transient expression of lacZ as determined by the MUG
assay in individual tilapia embryos. Since fertilised eggs resulted from eggs of wild type female fertilised with sperm of C58 male, only 50% of embryos will be
transgenic and will express lacZ. Four batches of embryos are scored, the first lot grown from uninjected eggs, batch 2 injected with sense-strand RNA, batch 3
with antisense strand RNA and batch 4 with ds RNA, the RNA being directed against the lacZ message. N ¼ 25, 21, 16 and 15, respectively. There is no
evidence of RNA interference with the lacZ expression.

To date only preliminary results are available. It is able. Expression of the tilapia serine 8 GnRH promoter has
obvious that these constructs cannot readily be tested in been tested with a reporter gene in spleen cells. Also tran-
tissue culture since the specific brain cell lines are not avail- sient expression in embryos is uninformative since the rele-

Fig. 6. Diagram giving details of the tilapia serine 8 GnRH decapeptide sequence and adjacent regions isolated in the Southampton laboratory.
 N. Maclean et al. / Gene 295 (2002) 265–277 275

vant messages are not expressed until later in development. VII cDNA driven by a CMV promoter, we have been able
Assay for effectivity is also tricky, since all fish will be to demonstrate transient expression of the protein in trans-
hemizygous and the first generation will be hemizygous genic zebrafish and catfish, and to show that the expressed
mosaics, and so complete effectivity may not be apparent protein does indeed clot human cells ‘in vitro’ (Muller et al.,
until non-mosaic homozygous transgenics can be produced. unpublished). Through a collaboration with Aquagene Inc.
It is significant that the transgenics produced with of Alachua, FL, USA we are now in process of producing
construct (b) show greatly reduced fertility and in some lines of transgenic tilapia expressing human factor VII
cases total sterility, and so there is some evidence here of driven by a tilapia tissue specific promoter and purifying
antisense activity. It is possible that homozygous fish for the protein from the tilapia tissues. Germ line transmission
this construct will prove entirely sterile. of the construct has already been achieved and we are now
testing the quality and quantity of protein produced.
3.4.6. Gene knock-down via ribozyme The particular attractions of the tilapia model in this
Ribozymes are RNA molecules with catalytic activity regard are: (a) comparatively low cost; (b) short generation
capable of facilitating the cutting of specific messages at time and therefore a fairly short timeframe between choice
target sites. They occur naturally in association with self of product and first production; (c) comparative ease of
splicing group I introns in cells such as Tetrahymena making and producing transgenic tilapia; and (d) public
(Lewin, 2000). Several types of ribozymes have been perception of fish as ‘lower’ vertebrates and thus more
described to show endoribonuclease activity. We have acceptable as biofactory systems than birds or mammals.
decided to use the hammerhead ribozyme type because it
is the most commonly used in mammals. Hammerhead ribo- 3.6. Future exploitation of the tilapia genome
zymes cut their targets at sites which are triplets of the Both on grounds of consumer acceptance of products and
sequence NHH, where N can be any base and H can be A, effectiveness of regulatory sequences in terms of compara-
C or T, but not G. However, all sequences are not cut with tive homology, it is clearly advantageous to isolate and use
equal efficiency. The sequence that is most sensitive to sequences of tilapia origin whenever possible. Although the
ribozyme hydrolysis is GUC (guanine, uridine, cytosine mining of the tilapia genome is at a very early stage, the
triplet), and this is also the only target sequence found so following promoter sequences of tilapia origin are already in
far to be involved with naturally occurring hammerhead use in our laboratory.
ribozyme activities. One of the main restrictions in using
ribozymes is the determination of the best target site on the Promoters:
messenger to knock out. In the case of the serine 8 GnRH H3 histone;
sequence this difficulty is somewhat simplified since only serine 8 GnRH;
five GUC codons are present in the sequence of the messen- L18 ribosomal protein (obtained from Molina et al.
ger so reducing the number of potential targets to be tested. Liege);
We have prepared ribozymes against all five of these sites. HSP 70;
There is also evidence that promoters specific to type III beta actin;
RNA polymerase are best, and the one being tested in our vitellogenin (obtained from Aquagene);
laboratory is a U6 promoter. U6 small nuclear RNA.

3.5. Future exploitation of transgenic tilapia With all of these promoters it is desirable to optimise the
length of upstream sequence in order to maximise expres-
A number of attractive avenues for future work are being sion in a construct. This optimisation can be difficult if a
explored. One of these is the production of growth enhanced relevant cell line is not available. With the serine 8 type
lines of transgenic tilapia which are transgenic only with GnRH promoter, there is the fortunate observation that
respect to gene sequences of tilapia origin, i.e. autotrans- this species of GnRH is expressed in spleen as well as in
genics. The construct in use is shown in Fig. 3. The coin- brain hypothallamus (White and Fernald, 1998). We have
jection strategy used in our existing growth enhanced fish therefore used primary cell cultures of tilapia spleen to test
(Rahman et al., 1997) will also be avoided. different lengths of this promoter when in combination with
a lacZ reporter. The expression obtained is specific to spleen
3.5.1. Use of transgenic tilapia as biofactories for cells. Fig. 6 shows some detail of the sequence so far
production of human pharmaceuticals isolated.
The exploitation of transgenic organisms for the produc- Tilapia coding sequences so far isolated include:
tion of valuable vaccines and other pharmaceutical products
is now far advanced, and examples include work with sheep, growth hormone (GH) (Liege laboratory);
pigs, goats and poultry. For some years we have been serine 8 GnRH;
exploring the benefits of transgenic tilapia to this end. tryptophan 7 GnRH;
Using a construct consisting of a human clotting factor LH.
276 N. Maclean et al. / Gene 295 (2002) 265–277

1996. Growth enhancement in transgenic tilapia by ectopic expression

Acknowledgements
of tilapia growth hormone. Mol. Mar. Biol. Biotechnol. 5, 62–70.
Misquitta, L., Paterson, B.M., 1999. Targeted disruption of gene function in
We are grateful for financial support to the DFID Fish Drosophila by RNA interference (RNA-i): A role of nautilus in
Genetics Programme for the benefit of developing countries, embryonic somatic muscle formation. Proc. Natl. Acad. Sci. USA 96,
to the EU for collaborative support, to the Government of 1451–1456.
Molina, A., Biemar, F., Müller, F., Iyengar, A., Prunet, P., Maclean, N.,
Iran for a studentship to Hamid Farahmand and to Aquagene
Martial, J.A., Muller, M., 2000. Cloning and expression analysis of a
Inc. of Alachua, FL. tilapia (Oreochromis mossambicus) HSP70 gene. FEBS Lett. 474, 5–
10.
Molina, A., Iyengar, A., Marins, L.F., Biemar, F., Hanley, S., Maclean, N.,
References Smith, T., Martial, J.A., Muller, M., 2001. Gene structure and promoter
function of a teleost ribosomal protein: a tilapia (Oreochromis mossam-
Beardmore, J., 1997. Transgenics: autotransgenics and allotransgenics. bicus) L18 gene. Biophys. Biochim. Acta 1520, 195–202.
Transgenic Res. 6, 107–108. Montoliu, L., Chávez, S., Vidal, M., 2000. Variegation associated with lacZ
Branch, A.D., 1998. A good antisense molecule is hard to find. Trends in transgenic animals: a warning note. Transgenic Res. 9, 237–239.
Biochem. Sci. 23, 45–50. Nam, Y.K., Cho, Y.S., Kim, D.S., 2000. Isogenic transgenic homozygous
Brem, G., Brenig, B., Hörstgen-Schwark, G., Winnacker, E.L., 1988. Gene fish induced by artificial parthenogenesis. Transgenic Res. 9, 463–469.
transfer in tilapia (Oreochromis niloticus). Aquaculture 68, 209–219. Palmiter, R.D., Brinster, R.L., Hammer, R.E., Trumbauer, M.E., Rosenfeld,
Cameron, E.R., Harvey, M.J.A., Onions, D.E., 1994. Transgenic science. M.G., Birnberg, N.C., Evans, R.M., 1982. Dramatic growth of mice that
Br. Vet. J. 150, 9–24. develop from eggs micro-injected with metallothionein-growth
De la Fuente, J., Guillén, I., Martı́nez, R., Estrada, M.P., 1999. Growth hormone fusion gene. Nature 300, 611–615.
regulation and enhancement in tilapia: basic research findings and Purdom, C.E., 1993. Genetics and Fish Breeding, Chapman & Hall,
their applications. Genet. Anal. Biomol. Eng. 15, 85–90. London.
Devlin, R.H., Vesaki, T.Y., Biagi, C.A., Donaldson, E.M., Swanson, P., Rahman, M.A., Maclean, N., 1992. Production of transgenic tilapia (Oreo-
Chan, W.-K., 1994. Extraordinary salmon growth. Nature 371, 209– chromis niloticus) by one-cell-stage micro-injection. Aquaculture 105,
210. 219–232.
Du, S.J., Gong, Z., Fletcher, G.L., Shears, M.A., King, M.J., Idler, D.R., Rahman, M.A., Maclean, N., 1999. Growth performance of transgenic
Hew, C.L., 1992. Growth enhancement in transgenic Atlantic salmon tilapia containing an exogenous piscine growth hormone gene. Aqua-
by the use of an “all fish” chimeric growth hormone gene construct. culture 173, 333–346.
Biotechnology 10, 179–187. Rahman, M.A., Iyengar, A., Maclean, N., 1997. Coinjection strategy
Dunham, R.A., Eash, J., Askins, J., Townes, T.M., 1987. Transfer of the improves integration efficiency and germline transmission of a growth
metallothionein-human growth hormone fusion gene into channel hormone gene construct in tilapia (Oreochromis niloticus). Transgenic
catfish. Trans. Am. Fish. Soc. 116, 87–91. Res. 6, 369–378.
FAO, 1999. The State of World Fisheries and Aquaculture, Food and Agri- Rahman, M.A., Mak, R., Ayad, H., Smith, A., Maclean, N., 1998. Expres-
culture Organization, Rome, Italy. sion of a novel piscine growth hormone gene results in growth enhance-
Fire, A., Xu, S., Montgomery, M.K., Kostas, S.A., Driver, S.E., Mello, ment in transgenic tilapia (Oreochromis niloticus). Transgenic Res. 7,
C.C., 1998. Potent and specific genetic interference by double-stranded 357–370.
RNA in Caenorhabditis elegans. Nature 391, 806–811. Rahman, M.A., Hwang, G., Razak, S.A., Sohm, F., Maclean, N., 2000.
Guise, K.S., Hackett, P.B., Faras, A.J., 1992. Transfer of genes encoding Copy number dependent transgene expression in hemizygous and
neomycin resistance, chloramphenicol acetyl transferase and growth homozygous transgenic tilapia (Oreochromis niloticus). Transgenic
hormone into goldfish and northern pike. In: Hew, C.L., Fletcher, Res. 9, 417–427.
G.L. (Eds.). Transgenic Fish, World Scientific, Singapore, pp. 142–163. Rahman, M.A., Ronyai, A., Engidaw, B.Z., Jauncey, K., Hwang, G., Smith,
Guyomard, R., Chourrout, D., Leroux, C., Houdebine, L.M., Pourrain, F., A., Roderick, E., Penman, D., Varadi, L., Maclean, N., 2001. Growth
1989. Integration and germ line transmission of foreign genes micro- and nutritional trials of transgenic Nile tilapia containing an exogenous
injected into fertilised trout eggs. Biochimie 71, 857–883. fish growth hormone gene. J. Fish Biol. 59, 62–78.
Hong, Y., Winkler, C., Schartl, M., 1996. Pluripotency and differentiation Razak, S.A., Hwang, G.-L., Rahman, M.A., Maclean, N., 1999. Growth
of embryonic stem cell lines from the medakafish (Oryzias latipes). performance and gonadal development of growth enhanced transgenic
Mech. Dev. 60, 33–44. tilapia Oreochromis niloticus (L.). following heat-shock-induced
Hong, Y., Winkler, C., Schartl, M., 1998. Production of medakafish triploidy. Mar. Biotechnol. 1, 533–544.
chimeras from a stable embryonic stem cell line. Proc. Natl. Acad. Ross, L.G., 2000. Environmental physiology and energetics. In: Beveridge,
Sci. USA 95, 3679–3684. M.C.M., McAndrew, B.J. (Eds.). Tilapias: Biology and Exploitation,
Hong, Y., Chen, S., Schartl, M., 2000. Embryonic stem cells in fish: current Kluwer Academic Publishers, Dordrecht, pp. 89–128.
status and perspectives. Fish Physiol. Biochem. 22 (2), 165–170. Seeburg, P.H., Mason, A.J., Stewart, T.A., Nikolics, R., 1987. The mamma-
Hussain, M.G., Chatterji, A., McAndrew, B.J., Johnstone, R., 1991. lian GnRH gene and its pivotal role in reproduction. Recent Prog.
Triploidy induction in Nile tilapia, Oreochromis niloticus, L. using Horm. Res. 43, 69–98.
pressure, heat and cold shocks. Theor. Appl. Genet 81, 6–12. Shears, M.A., Fletcher, G.L., Hew, C.L., Gauthier, S., Davies, P.L., 1991.
Iyengar, A., Müller, F., Maclean, N., 1996. Regulation and expression of Transfer, expression and stable inheritance of antifreeze protein genes
transgenes in fish – a review. Transgenic Res. 5, 147–166. in Atlantic Salmon (Salmo salar). Mol. Mar. Biol. Biotechnol. 1, 58–63.
Lewin, B., 2000. Genes VII, Oxford University Press, Oxford. Sigurdsson, S., Trujillo, K., Song, B., Stratton, S., Sung, P., 2001. Basis for
Maclean, N., 1998. Regulation and exploitation of transgenes in fish. Mutat. avid homologous DNA strand exchange by human Rad51 and RPA. J.
Res. 399, 255–266. Biol. Chem. 276, 8798–8806.
Maclean, N., Laight, R., 2000. Transgenic fish: an evaluation of benefits Sokol, D.L., Murray, J.D., 1996. Antisense and ribozyme construct in
and risks. Fish Fish. 1, 146–172. transgenic animals. Transgenic Res. 5, 363–371.
Martinez, R., Estrada, M.P., Berlanga, J., Guillén, I., Hernández, O., Stuart, G.W., Vielkind, J.R., McMurray, J.V., Westerfield, M., 1990. Stable
Cabrera, E., Pimental, R., Morales, R., Herrera, F., Morales, A., Piña, lines of transgenic zebrafish exhibit reproducible patterns of transgene
J.C., Abad, Z., Sánchez, V., Melamed, P., Lleonart, R., de la Fuente, J., expression. Development 109, 577–584.
 N. Maclean et al. / Gene 295 (2002) 265–277 277

Sun, L., Bradford, C.S., Ghosh, C., Collodi, P., Barnes, D.W., 1995. ES-like Yoon, S.J., Hallerman, E.M., Gross, M.L., Liu, Z., Schneider, J.F., Faras,
cell cultures derived from early zebrafish embryos. Mol. Mar. Biol. A.J., Hackett, P.B., Kapuscinski, A.R., Guise, K.S., 1990. Transfer of
Biotechnol. 4, 193–199. the gene for neomycin resistance into goldfish (Carassius auratus).
Volckaert, F.A., Hellemans, B.A., Galbusera, P., Sekkali, B., Belayew, A., Aquaculture 85, 21–33.
Ollevier, F., 1994. Replication, expression and fate of foreign DNA Zhang, P., Hayat, M., Joyce, C., Gonzalez-Villasenor, L.I., Lin, C.M.,
during embryonic and larval development of the African catfish (Clar- Dunham, R.A., Chen, T.T., Powers, D.A., 1990. Gene transfer,
ias gariepinus). Mol. Mar. Biol. Biotechnol. 3, 57–69. expression and inheritance of pRSV-Rainbow Trout-H cDNA in the
White, R.B., Fernald, R.D., 1998. Genomic structure and expression sites of common carp Cyprinus carpio (Linnaeus). Mol. Reprod. Dev. 25, 3–
three gonadotropin-releasing hormone genes in one species. Gen. 13.
Comp. Endocrinol. 112, 17–25. Zhao, Z., Cao, Y., Li, M., Meng, A., 2001. Double-stranded RNA injection
Wianny, F., Zernicka-Goetz, M., 2000. Specific interference with gene produces non-specific defects in zebrafish. Dev. Biol. 229, 215–223.
function by double-stranded RNA in early mouse development. Nat.
Cell. Biol. 2, 70–75.