- Email: [email protected]
Knockout of one acetylcholinesterase allele in the mouse
Chemico-Biological Interactions 119 – 120 (1999) 289 – 299
Knockout of one acetylcholinesterase allele in the mouse Weihua Xie a, Phillip J. Wilder a, Judith Stribley b, Arnaud Chatonnet c, Angie Rizzino a, Palmer Taylor d, Steven H. Hinrichs b, Oksana Lockridge a,* a
Eppley Institute, Uni6ersity of Nebraska Medical Center, 600 S. 42nd St., Omaha, NE 68198 -6805, USA b Department of Pathology and Microbiology, Uni6ersity of Nebraska Medical Center, Omaha, NE 68198 -6805, USA c INRA, Montpellier, France d Department of Pharmacology, Uni6ersity of California, San Diego, USA
Abstract One allele of the AChE gene (ACHE) was knocked out in embryonic stem (ES) cells by homologous recombination. The targeting vector contained 2 kb of a TK gene cassette for negative selection, 884 bp of ACHE including exon 1, 1.6 kb of a Neor gene cassette for positive selection, 5.2 kb of the ACHE Bam HI fragment including exon 6, and 3 kb of Bluescript. The use of this vector deleted exons 2 – 5, which removed 93% of the ACHE coding sequence including the signal peptide, the active site serine, and the histidine and glutamic acid of the catalytic triad. The gene targeting vector was transfected into ES cells by electroporation. Colonies resistant to G418 and gancyclovir were screened for homologous recombination by Southern blotting. Out of 200 colonies, four were found to have undergone homologous recombination. These four ACHE ( 9) ES cell lines were expanded to provide cells for microinjection into C57Bl/6 mouse blastocysts. The injected blastocysts were implanted into pseudopregnant CD/1 white mice. More than 200 injected blastocysts were transferred into 20 mice. More than 65 mice were born, of which 11 were chimeras. Chimeras were identified by their black and agouti coat color. Littermates were all black. Thus far, seven male chimeras have been bred with more than 130 C57Bl/6 females to generate 26 agouti mice out of 199 living offspring. This demonstrated that the ACHE (9 )
* Corresponding author. Tel: +1-402-559-6032; fax: + 1-402-559-4651. E-mail address: [email protected] (O. Lockridge) 0009-2797/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 0 0 9 - 2 7 9 7 ( 9 9 ) 0 0 0 3 9 - 3
290
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
ES cells contributed to the germline. Offspring with agouti coat color have a 50% chance of carrying the knockout allele. The 26 agouti offspring were screened for an ACHE ( 9 ) genotype by tail biopsy PCR. Ten out of 26 agouti mice are heterozygous ACHE knockout mice, and they are healthy and alive at 29 days of age. We expect a phenotype to appear in nullizygous animals. © 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Acetylcholinesterase (AChE); AChE gene (ACHE); Knockout mouse
1. Introduction Acetylcholinesterase (AChE) has a function at cholinergic synapses where it terminates nerve impulse transmission by hydrolyzing acetylcholine. However, AChE also appears to have nonsynaptic functions during development [1–5]. To identify this function, we are making the ACHE knockout mouse. As a first step, we made the heterozygous knockout mouse in which only one functional ACHE allele is present. It has generally been assumed that the AChE enzyme is so important for life that mutations leading to loss of activity are incompatible with life. It was not known whether a mammal could live if one of its ACHE alleles was nonfunctional. In this report, we demonstrate that deletion of one ACHE allele has no obvious deleterious effects. This work is still in progress. We are planning to produce the nullizygote, that is a mouse with no AChE enzyme activity, by breeding heterozygotes.
2. Methods
2.1. Construction of the 12.7 kb mouse ACHE knockout gene targeting 6ector A lambda FIX II clone of ACHE from mouse strain 129 was isolated from a library purchased from Stratagene. The lambda FIX II clone was 13.6 kb in size and included the entire ACHE gene (Fig. 1). Fig. 2 shows the gene targeting vector. To construct the short arm of the vector, we first cloned an Nhe I fragment from mouse ACHE into the plasmid Litmus 39. This fragment includes exon 1 and part of intron 1. The 1067 bp in this fragment were sequenced. The sequence allowed us to design primers for PCR amplification that would yield an 884 bp fragment that included all of exon 1 but excluded all of intron 1. Xho I cloning sites were added to the primers. The 884 bp PCR fragment was cloned and resequenced. After it was confirmed that there were no errors in the sequence, the 884 bp of exon 1 were cloned into plasmid MJK-KO. This gene targeting plasmid has a 2 kb TK gene cassette for negative selection to reduce random insertion and a 1.6 kb Neor gene cassette for positive selection to increase the targeting frequency. The backbone of plasmid MJK-KO vector has 3 kb of Bluescript SK + . The correct orientation of the 884 bp of mouse ACHE was verified by sequencing.
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
291
To construct the long arm of the vector, Dr. A. Chatonnet cloned the Bam HI fragment of mouse ACHE into pUC. The 5.2 kb Bam HI fragment includes exon 6. This fragment was moved out of pUC and into the gene targeting vector, which already contained the short arm of ACHE. The short and long arms of ACHE in the targeting vector were oriented in the same direction as verified by PCR, by restriction enzyme mapping and by sequencing. The TK gene and the Neor gene were oriented opposite to the direction of the two ACHE fragments. Use of this targeting vector deletes 5 kb of the ACHE gene encoding the signal peptide, the catalytic triad, and amino acids 1-543, which represents 93% of the total 573 amino acids in AChE (Fig. 1).
2.2. Introduction of the mouse ACHE knockout gene into ES cells The 12.7 kb targeting vector was linearized with Not I in preparation for transfection. The linearized targeting vector was extracted with phenol-chloroform and sterilized by ethanol precipitation. The linearized targeting vector was dissolved in sterile TE buffer to a final DNA concentration of 2 mg/ml. Twenty micrograms of the linearized, sterile plasmid was mixed with 30× 106 R1 ES cells [6,7] in 0.8 ml of 1 × PBS. The cells were electroporated at 240 V, 500 mF using a Gene Pulser Electroporator (BioRad). The electroporated ES cells were transferred into six 10-cm dishes containing a feeder layer of neomycin resistant mouse embryonic fibroblasts that had been gamma-irradiated with a total dosage of 5000 Rad. The dish also contained 10 ml of medium (DMEM, 15% FBS, 0.1 mM b-mercaptoethanol, 2 ng/ml LIF). The next day, G418 at a concentration of 250 mg/ml was added to the medium, and 2 days after electroporation 2 mM gancyclovir was added. Single colonies resistant to G418 and gancyclovir were transferred into 96 well plates and expanded into 24 well plates to grow enough ES cells for
Fig. 1. Structure of the mouse ACHE gene. Black boxes are 6 exons. The white box between exons 5 and 6 is a retained intron that is present in the mRNA. Exon 1 has only untranslated sequences. The protein is encoded by exons 2–6. The 5 kb deleted from the gene to make the knockout mouse are indicated. The knockout ACHE allele has no active site and no catalytic triad.
292
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
Fig. 2. ACHE knockout gene targeting vector. The short arm has 884 bp of ACHE from the exon 1 region of the gene. The long arm has 5.2 kb of ACHE from the 3% end of the ACHE gene and includes exon 6. The Neo and TK genes are oriented opposite to the ACHE short and long arms. The total size of the gene targeting vector is 12.7 kb.
screening. A total of 200 single colonies were screened for homologous recombination by Southern blotting [8]. The genomic DNA was digested with Bam HI. The probe was outside of the targeted region so that the hybridization would detect homologous recombination events rather than random insertion. All the screening was completed within 7 days. This short time meant that ES cells divided only seven times before they were frozen in liquid nitrogen. It was important to keep the number of cell divisions to a minimum to maintain totipotency. Cell lines identified as positive were expanded in T75 flasks and immediately frozen in liquid nitrogen, at a density of 4×106 cells per vial. To provide material for further testing, positive ES cells were grown without a feeder layer.
2.3. Chromosome analysis of three ACHE (9) ES cell lines Chromosome analysis [9] was performed by the Cytogenetics Laboratory of the University of Nebraska Medical Center. Cell lines 3D6, 7D1, 8A1 were analyzed, but not cell line 4A5 because this cell line appeared to contain a significant proportion of differentiated cells. Cell line 4A5 was not injected into blastocysts.
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
293
Cultures of all three cell lines were observed for appropriate growth and then subcultured onto coverslips. Cultures were harvested following 30 min of exposure to Colcemid, exposure to 0.75 M KCl hypotonic solution, and fixation with a 3:1 methanol:glacial acetic acid mixture. Slides were prepared and G-banded using Wright’s stain. With a computer analysis system, a minimum of 20 metaphases from each cell line were digitized and printed. Counts and analysis were performed on printed images of each cell.
2.4. Microinjection of ES cells into blastocysts The targeted ES cells were grown in T25 flasks on gamma-irradiated mouse embryonic fibroblasts in preparation for injection into blastocysts [10,11]. Two days before microinjection, 3× 106 cells were seeded into a T25 gelatin-coated flask without feeder cells. Blastocysts were harvested from C57Bl/6 females and microinjected with 10 – 15 ACHE (9 ) ES cells per blastocyst. Eight–ten blastocysts were surgically transferred into one side of the uterus of a pseudopregnant CD1 white female. Gestation in mice is 21 days. After the pups were born, mice with the highest proportion of agouti coat color were expected to have the highest proportion of ES cells in the germline.
3. Results
3.1. Homologous recombination in ES cells Out of 200 colonies, four cell lines 3D6, 4A5, 7D1, and 8A1 were found to have undergone homologous recombination. The four positive clones were identified by Southern blotting using a probe outside of the targeted region. Positive clones had a 5.0 kb Bam HI band for the knocked out allele, and a 6.5 kb band for the wild-type allele (Fig. 3). To confirm that there was no random insertion in these four positive cell lines we probed Southern blots with a fragment of the Neor gene on DNA digested with Xba I, Xho I, and Nhe I. The blots were hybridized with three different probes both inside and outside the targeted region. In addition, we used PCR to confirm that homologous recombination had taken place. The sense primer was upstream of exon 1, outside of the targeted region, and the antisense primer was in the Neor gene (data not shown).
3.2. Chromosome analysis Table 1 shows the results from chromosome analysis of the three positive ES cell lines that were injected into blastocysts. Cells with chromosome counts greater or equal to 36 were cytogenetically analyzed for the loss of the Y chromosome. All three cell lines had cells that were missing the Y chromosome: 3D6 had four out of 20 (20%), 7D1 had two (10%), and 8A1 had 16 (80%). The 7D1 cell line had one cell with what appeared to be a metaphase chromosome (centric fusion). 3D6 and
294
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
8A1 cell lines had cells with chromosome breaks and gaps. None of the cell lines showed homogeneously staining regions (HSRs), double minutes (DMs), or ring chromosomes. On the basis of the cytogenetic analysis, cell line 7D1 appeared to be the best cell line. However, this cell line yielded no chimeras. Only cell line 3D6 yielded chimeras with the ACHE knockout in the germline, a result that could not have been predicted from the chromosome analysis.
3.3. Generation of the ACHE (9) mouse From the first microinjection trial, one male chimera with 50% agouti coat color was born from the cell line 8A1. This chimera produced 100 black mice but no agouti mice. It was concluded that the ACHE knockout was not in the germline of this chimera. From the second microinjection trial, two male chimeras (3-1-I, 3-1-II) with 100% agouti coat color and one male chimera (3-1-III) with 75% agouti coat color were born from the ACHE (9 ) ES cell line 3D6. One male chimera (8-1-I) with 100% agouti coat color was obtained with the cell line 8A1.
Fig. 3. Southern blot showing one ACHE knockout allele. The 6.0 kb band is the wild-type allele. The 5.6 kb band is the knocked-out allele. The probe was outside of the targeted region
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
295
Table 1 Chromosome analysis of three ACHE ( 9) ES cell linesa Cell lines
3D6
7D1
8A1
Number of chromosomes out of 20 metaphase cells examined c35 c36 c37 c38 c39 c40 c41 c42 Modal chromosome number Number of metaphase analyzed Number of cells with (2n) modal number Percentage of cells with 2n (%) Number of cells with chromosome count \36 and minus the Y chromosome Percentage of cells \36 and minus Y chromosome (%) Percentage of metacentric chromosome (%) Presence of hypomodal cells Presence of hypermodal cells Presence of homogeneously staining regions (HSRs) Presence of double minutes (DMs) Presence of ring chromosomes Percentage of breaks and gaps (%)
0 0 0 0 8 11 1 0 40 20 11 55 4 20 0 yes yes no no no 10
0 0 0 1 0 17 2 0 40 20 17 85 2 10 5 yes yes no no no 0
1 0 0 3 13 2 1 0 39 20 2 10 16 80 0 yes yes no no no 20
a
Only cell line 3D6 produced chimeras that had the ACHE knockout allele in the germline.
All four chimeras have been bred with more than 110 C57Bl/6 females. Table 2 lists the offspring of four chimeras from the 2nd ES injection trial. To date, a total of 26 agouti mice and 73 black mice have been born. All 26 agouti coat colored mice are from chimeras 3-1-I and 3-1-III. This shows that cell line 3D6 yielded chimeras with germline transmission of the ACHE (9 ) genotype. Use of cell line 8A1 did not result in germline transmission of the ES cell genome, though it did produce chimeras. The chimera 3-1-II with 100% agouti coat color appeared to be sterile, because it has had no offspring to date. Offspring with agouti coat color have a 50% chance of carrying the ACHE knockout allele (9). Twenty-six agouti offspring were screened for ACHE (9 ) by PCR of DNA extracted from tails. Twelve out of 26 agouti mice were heterozygous ACHE knockout mice. This was determined by PCR analysis using a sense primer in the mouse ACHE gene and an antisense primer in the Neor gene. Ten heterozygous knockouts are still healthy and alive at age 29 days. Male heterozygous ACHE knockout mice and female heterozygous ACHE knockout mice will be mated to get homozygous ACHE knockout mice. From the third microinjection trial, one male chimera 3-456 with 100% agouti coat color was born from cell line 3D6. One male chimera 8-458 with 100% agouti coat color was born from cell line 8A1. Four female chimeras 8-457, 8-459, 8-460,
296
Table 2 Coat color for offspring from the mating of four male chimeras with 110 C57 Bl/6a femalesb,c C57Bl/6 female mates
Litter
Pups D.O.B
Total pups
Born dead
Born alive
Survive21 days
Black mice
Agouti mice
PCR screen agouti mice ACHE ( 9)
3-1-I 11/24/97
29
1 1 1 2
01/29/98 03/12/98 03/15/98 03/22/98
7 13 10 16
0 2 1 0
7 11 9 16
0 11 4 11
0 1 1 0
0 10 3 11
N/D 4 3 3
3-1-II 11/24/97
29
0
0
0
0
0
0
0
0
N/D
3-1-III 11/28/97
27
1 1 1 1 1 1 1
02/02/98 02/12/98 02/15/98 02/23/98 03/02/98 03/12/98 03/15/98
10 4 5 5 6 9 10
0 4 0 5 6 1 0
10 0 5 0 0 8 0
8 0 0 0 0 8 0
8 0 0 0 0 6 0
0 0 0 0 0 2 0
0 N/D N/D 2 (dead) N/D 0 N/D
8-I 12/01/97
25
1 1 1 1 2 2 2 1
02/06/98 2/12/98 2/23/98 3/5/98 3/12/98 3/20/98 3/24/98 4/1/98
8 7 7 6 20 15 10 12
0 2 2 0 4 0 0 2
8 5 5 6 16 15 10 10
8 5 0 0 16 15 10 3
8 5 0 0 16 15 10 3
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0
Total
110
23
180
29
151
99
73
26
12
a
C57 black/6 mice bought from Jackson Laboratory Inc., Bar Harbor, Maine, (207)-288-5845. N/D, not done. c The male chimeras were produced in the 2nd ES injection trial. b
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
Male, chimera D.O.B.
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
297
and 8-461 with 100% agouti coat color were born from 8A1 ACHE (9 ) ES cell line. The chromosome analysis results showed that 80% of cells in cell line 8A1 had lost the Y chromosome, which explains why this cell line produced four female chimeras. Table 3 summarizes the results from 480 blastocysts injected with three ES cell lines. Eleven chimeras were produced, of which seven were males and four were females. Two of the male chimeras transmitted the ACHE knockout in their germline.
4. Discussion Genes with a high GC content can have a low rate of homologous recombination. Therefore, it was a surprise that the mouse ACHE gene underwent homologous recombination with the high frequency of 2%. Of the 11 chimeras produced in this study, only two transmitted the ACHE knockout allele in their germline. Our finding that mice with only one ACHE allele survive to birth and adulthood is novel, because no mammalian species has previously been found to have a null ACHE allele. The wobbler mouse has 50% of normal ACHE mRNA levels [12], but the reduced level is a consequence of a mutation on chromosome 11 rather than at the ACHE locus on chromosome 5. Examples of null ACHE alleles in living animals have been reported in the Drosophila fly and the C. elegans worm. These animals survive when a single ACHE allele is deleted. However, mutations at the ACHE locus that abolish AChE activity in these animals are embryonic lethals [13,14]. The fact that mice with one null ACHE allele are healthy suggests that people may also be healthy when they have only one functional ACHE allele. There are two case reports of genetically inherited, low AChE activity in healthy families [15,16]. It is possible that these families have one deficient ACHE allele. There is no human disease of AChE deficiency, though there are two diseases where the AChE anchors are deficient, namely paroxysmal nocturnal hemoglobinuria [17] and familial myasthenia [18,19]. The absence of diseases of AChE deficiency, the importance of AChE for breathing, and the known lethality of nerve agents that inhibit AChE, have led to the expectation that the homozygous Table 3 Eleven chimeras from 480 injected blastocysts ACHE ( 9 ) cell lines
Chimeras
Males
Females
Germline transmission
3D6 4A5 7D1 8A1 Total
4 not used 0 7 11
4
0
2
0 3 7
0 4 4
0 0a 2
a
Female chimeras were not tested for germline transmission.
298
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
knockout mouse will be unable to live, and might die soon after AChE is first expressed in the mouse embryo on day 9 [20]. However, if contrary to expectation, the nullizygote does live, then it will be interesting to determine what compensatory enzymes perform the function of AChE. A prime candidate for this role is butyrylcholinesterase. During early development of the rat neural tube, butyrylcholinesterase and AChE enzymes are found in the same cells [21]. Therefore, it is possible that butyrylcholinesterase could substitute for AChE in some functions.
Acknowledgements We thank Dr. S. Steven Potter for the MJK-KO plasmid. A. Chatonnet isolated the lambda clone and prepared clones in pUC; the idea to make the knockout mouse came from Dr. Chatonnet. P. Taylor provided the mouse ACHE sequence. P. Wilder and A. Rizzino of the gene targeting facility at UNMC did the ES cell work. J. Stribley and S.H. Hinrichs of the microinjection facility at UNMC transferred the ES cells into mouse blastocysts, implanted the blastocysts, and arranged the mouse breeding. W. Xie made the gene targeting vector, screened ES colonies with Southern blots, arranged mouse breeding, and screened animals by PCR and other methods. O. Lockridge coordinated the project. Supported by U.S. Army Medical Research and Materiel Command DAMD 17-94-J-4005 and DAMD 17-97-1-7349 (to O.L.), NIH grant R01 DA011707 (to O.L.), Association Francaise Contre les Myopathies (MNM1997) (to A.C.), Nebraska State Research Initiative (to S.H.H. and A.R.) and UNMC seed grant c 98-005 (to O.L.). The opinions or assertions contained herein belong to the authors and should not be construed as the official views of the U.S. Army or the Department of Defense.
References [1] P.G. Layer, Nonclassical roles of cholinesterases in the embryonic brain and possible links to Alzheimer Disease, Alzheimer Dis. Assoc. Dis. 9 (Suppl 2) (1995) 29 – 36. [2] P.G. Layer, Non-classical actions of cholinesterases: role in cellular differentiation, tumorigenesis, and Alzheimer’s disease, Neurochem. Int. 28 (1996) 491 – 495. [3] P.G. Layer, T. Weikert, R. Alber, Cholinesterases regulate neurite growth in chick nerve cells in vitro by means of a non-enzymatic mechanism, Cell Tiss. Res. 273 (1993) 219 – 226. [4] R.T. Robertson, J. Yu, Acetylcholinesterase and neural development: new tricks for an old dog?, News Physiol. Sci. 8 (1993) 266–272. [5] W.J. Wolfgang, M.A. Forte, Expression of acetylcholinesterase during system development in Drosophila, Dev. Biol. 131 (1989) 321 – 330. [6] M.L. Hooper, Embryonic Stem Cells-Introducing Planned Changes into the Animal Germline, Harwood Academic Publishers, Chur, Switzerland, 1994, pp. 56 – 81. [7] A. Nagy, J. Rossant, R. Nagy, W. Abramow-Newerly, J.C. Roder, Derivation of completely cell culture-derived mice from early-passage embryonic stem cells, Proc. Natl. Acad. Sci. 90 (1993) 8424–8428. [8] D.M. Glover, B.D. Hames, DNA Cloning 4: Mammalian Systems, Chapter 5. Oxford University Press, New York, 1995, pp.143–184.
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
299
[9] T.C. Hsu, K. Benirschke, An Atlas of Mammalian Chromosomes, 1, Folio 17. Springer-Verlage, New York, 1967. [10] E.J. Robertson, Teratocarcinomas and Embryonic Stem Cells — A Practical Approach, IRL Press, Oxford, 1987, pp. 108–151. [11] B. Hogan, R. Beddington, F. Costantini, E. Lacy, Manipulating the Mouse Embryo. A Laboratory Manual, Cold Spring Harbor Laboratory Press, New York, 1994, pp. 106 – 172. [12] R. Pernas-Alonso, A.E. Schaffner, C.T. Hansen, J.L. Barker, U. di Porzio, Acetylcholine esterase and peripherin mRNA level decrease in wobbler mouse, NeuroReport 6 (1995) 597 – 600. [13] J.C. Hall, D.R. Kankel, Genetics of acetylcholinesterase in Drosophila melanogaster, Genetics 83 (1976) 517–535. [14] C.D. Johnson, J.B. Rand, R.K. Herman, B.D. Stern, R.L. Russell, The acetylcholinesterase genes of C. elegans: identification of a third gene (ace-3) and mosaic mapping of a synthetic lethal phenotype, Neuron 1 (1988) 165–173. [15] R.J. Johns, Familial reduction in red-cell cholinesterase, New England J. Med. 267 (1962) 1344–1348. [16] K. Shinohara, K.R. Tanaka, Hereditary deficiency of erythrocyte acetylcholinesterase, Am. J. Hematol. 7 (1979) 313–321. [17] K. Nafa, P.J. Mason, P. Hillmen, L. Luzzatto, M. Bessler, Mutations in the PIG — A gene causing paroxysmal nocturnal hemoglobinuria are mainly of the frameshift type, Blood 96 (1995) 4650 – 4655. [18] D.O. Hutchinson, T.J. Walls, S. Nakano, S. Camp, P. Taylor, C.M. Harper, R.V. Groover, H.A. Person, D.G. Jamieson, A.G. Engel, Congenital endplate acetylcholinesterase deficiency, Brain 116 (1993) 633–653. [19] S. Camp, S. Bon, Y. Li, D.K. Getman, A.G. Engel, J. Massoulie´, P. Taylor, Patients with congenital myasthenia associated with end-plate acetylcholinesterase deficiency show normal sequence, mRNA splicing, and assembly of catalytic subunits, J. Clin. Invest. 95 (1995) 333 – 340. [20] N. Al-Fakhri, G. Bogusch, Activity of acetylcholinesterase and unspecific cholinesterase during differentiation of somites in mouse embryos, Anat. Embryol. 192 (1995) 257 – 264. [21] C. Koenigsberger, P. Hammond, S. Brimijoin, Developmental expression of acetyl- and butyrylcholinesterase in the rat: Enzyme and mRNA levels in embryonic dorsal root ganglia, Brain Res. 787 (1998) 248–258.
.
Knockout of one acetylcholinesterase allele in the mouse Weihua Xie a, Phillip J. Wilder a, Judith Stribley b, Arnaud Chatonnet c, Angie Rizzino a, Palmer Taylor d, Steven H. Hinrichs b, Oksana Lockridge a,* a
Eppley Institute, Uni6ersity of Nebraska Medical Center, 600 S. 42nd St., Omaha, NE 68198 -6805, USA b Department of Pathology and Microbiology, Uni6ersity of Nebraska Medical Center, Omaha, NE 68198 -6805, USA c INRA, Montpellier, France d Department of Pharmacology, Uni6ersity of California, San Diego, USA
Abstract One allele of the AChE gene (ACHE) was knocked out in embryonic stem (ES) cells by homologous recombination. The targeting vector contained 2 kb of a TK gene cassette for negative selection, 884 bp of ACHE including exon 1, 1.6 kb of a Neor gene cassette for positive selection, 5.2 kb of the ACHE Bam HI fragment including exon 6, and 3 kb of Bluescript. The use of this vector deleted exons 2 – 5, which removed 93% of the ACHE coding sequence including the signal peptide, the active site serine, and the histidine and glutamic acid of the catalytic triad. The gene targeting vector was transfected into ES cells by electroporation. Colonies resistant to G418 and gancyclovir were screened for homologous recombination by Southern blotting. Out of 200 colonies, four were found to have undergone homologous recombination. These four ACHE ( 9) ES cell lines were expanded to provide cells for microinjection into C57Bl/6 mouse blastocysts. The injected blastocysts were implanted into pseudopregnant CD/1 white mice. More than 200 injected blastocysts were transferred into 20 mice. More than 65 mice were born, of which 11 were chimeras. Chimeras were identified by their black and agouti coat color. Littermates were all black. Thus far, seven male chimeras have been bred with more than 130 C57Bl/6 females to generate 26 agouti mice out of 199 living offspring. This demonstrated that the ACHE (9 )
* Corresponding author. Tel: +1-402-559-6032; fax: + 1-402-559-4651. E-mail address: [email protected] (O. Lockridge) 0009-2797/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 0 0 9 - 2 7 9 7 ( 9 9 ) 0 0 0 3 9 - 3
290
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
ES cells contributed to the germline. Offspring with agouti coat color have a 50% chance of carrying the knockout allele. The 26 agouti offspring were screened for an ACHE ( 9 ) genotype by tail biopsy PCR. Ten out of 26 agouti mice are heterozygous ACHE knockout mice, and they are healthy and alive at 29 days of age. We expect a phenotype to appear in nullizygous animals. © 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Acetylcholinesterase (AChE); AChE gene (ACHE); Knockout mouse
1. Introduction Acetylcholinesterase (AChE) has a function at cholinergic synapses where it terminates nerve impulse transmission by hydrolyzing acetylcholine. However, AChE also appears to have nonsynaptic functions during development [1–5]. To identify this function, we are making the ACHE knockout mouse. As a first step, we made the heterozygous knockout mouse in which only one functional ACHE allele is present. It has generally been assumed that the AChE enzyme is so important for life that mutations leading to loss of activity are incompatible with life. It was not known whether a mammal could live if one of its ACHE alleles was nonfunctional. In this report, we demonstrate that deletion of one ACHE allele has no obvious deleterious effects. This work is still in progress. We are planning to produce the nullizygote, that is a mouse with no AChE enzyme activity, by breeding heterozygotes.
2. Methods
2.1. Construction of the 12.7 kb mouse ACHE knockout gene targeting 6ector A lambda FIX II clone of ACHE from mouse strain 129 was isolated from a library purchased from Stratagene. The lambda FIX II clone was 13.6 kb in size and included the entire ACHE gene (Fig. 1). Fig. 2 shows the gene targeting vector. To construct the short arm of the vector, we first cloned an Nhe I fragment from mouse ACHE into the plasmid Litmus 39. This fragment includes exon 1 and part of intron 1. The 1067 bp in this fragment were sequenced. The sequence allowed us to design primers for PCR amplification that would yield an 884 bp fragment that included all of exon 1 but excluded all of intron 1. Xho I cloning sites were added to the primers. The 884 bp PCR fragment was cloned and resequenced. After it was confirmed that there were no errors in the sequence, the 884 bp of exon 1 were cloned into plasmid MJK-KO. This gene targeting plasmid has a 2 kb TK gene cassette for negative selection to reduce random insertion and a 1.6 kb Neor gene cassette for positive selection to increase the targeting frequency. The backbone of plasmid MJK-KO vector has 3 kb of Bluescript SK + . The correct orientation of the 884 bp of mouse ACHE was verified by sequencing.
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
291
To construct the long arm of the vector, Dr. A. Chatonnet cloned the Bam HI fragment of mouse ACHE into pUC. The 5.2 kb Bam HI fragment includes exon 6. This fragment was moved out of pUC and into the gene targeting vector, which already contained the short arm of ACHE. The short and long arms of ACHE in the targeting vector were oriented in the same direction as verified by PCR, by restriction enzyme mapping and by sequencing. The TK gene and the Neor gene were oriented opposite to the direction of the two ACHE fragments. Use of this targeting vector deletes 5 kb of the ACHE gene encoding the signal peptide, the catalytic triad, and amino acids 1-543, which represents 93% of the total 573 amino acids in AChE (Fig. 1).
2.2. Introduction of the mouse ACHE knockout gene into ES cells The 12.7 kb targeting vector was linearized with Not I in preparation for transfection. The linearized targeting vector was extracted with phenol-chloroform and sterilized by ethanol precipitation. The linearized targeting vector was dissolved in sterile TE buffer to a final DNA concentration of 2 mg/ml. Twenty micrograms of the linearized, sterile plasmid was mixed with 30× 106 R1 ES cells [6,7] in 0.8 ml of 1 × PBS. The cells were electroporated at 240 V, 500 mF using a Gene Pulser Electroporator (BioRad). The electroporated ES cells were transferred into six 10-cm dishes containing a feeder layer of neomycin resistant mouse embryonic fibroblasts that had been gamma-irradiated with a total dosage of 5000 Rad. The dish also contained 10 ml of medium (DMEM, 15% FBS, 0.1 mM b-mercaptoethanol, 2 ng/ml LIF). The next day, G418 at a concentration of 250 mg/ml was added to the medium, and 2 days after electroporation 2 mM gancyclovir was added. Single colonies resistant to G418 and gancyclovir were transferred into 96 well plates and expanded into 24 well plates to grow enough ES cells for
Fig. 1. Structure of the mouse ACHE gene. Black boxes are 6 exons. The white box between exons 5 and 6 is a retained intron that is present in the mRNA. Exon 1 has only untranslated sequences. The protein is encoded by exons 2–6. The 5 kb deleted from the gene to make the knockout mouse are indicated. The knockout ACHE allele has no active site and no catalytic triad.
292
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
Fig. 2. ACHE knockout gene targeting vector. The short arm has 884 bp of ACHE from the exon 1 region of the gene. The long arm has 5.2 kb of ACHE from the 3% end of the ACHE gene and includes exon 6. The Neo and TK genes are oriented opposite to the ACHE short and long arms. The total size of the gene targeting vector is 12.7 kb.
screening. A total of 200 single colonies were screened for homologous recombination by Southern blotting [8]. The genomic DNA was digested with Bam HI. The probe was outside of the targeted region so that the hybridization would detect homologous recombination events rather than random insertion. All the screening was completed within 7 days. This short time meant that ES cells divided only seven times before they were frozen in liquid nitrogen. It was important to keep the number of cell divisions to a minimum to maintain totipotency. Cell lines identified as positive were expanded in T75 flasks and immediately frozen in liquid nitrogen, at a density of 4×106 cells per vial. To provide material for further testing, positive ES cells were grown without a feeder layer.
2.3. Chromosome analysis of three ACHE (9) ES cell lines Chromosome analysis [9] was performed by the Cytogenetics Laboratory of the University of Nebraska Medical Center. Cell lines 3D6, 7D1, 8A1 were analyzed, but not cell line 4A5 because this cell line appeared to contain a significant proportion of differentiated cells. Cell line 4A5 was not injected into blastocysts.
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
293
Cultures of all three cell lines were observed for appropriate growth and then subcultured onto coverslips. Cultures were harvested following 30 min of exposure to Colcemid, exposure to 0.75 M KCl hypotonic solution, and fixation with a 3:1 methanol:glacial acetic acid mixture. Slides were prepared and G-banded using Wright’s stain. With a computer analysis system, a minimum of 20 metaphases from each cell line were digitized and printed. Counts and analysis were performed on printed images of each cell.
2.4. Microinjection of ES cells into blastocysts The targeted ES cells were grown in T25 flasks on gamma-irradiated mouse embryonic fibroblasts in preparation for injection into blastocysts [10,11]. Two days before microinjection, 3× 106 cells were seeded into a T25 gelatin-coated flask without feeder cells. Blastocysts were harvested from C57Bl/6 females and microinjected with 10 – 15 ACHE (9 ) ES cells per blastocyst. Eight–ten blastocysts were surgically transferred into one side of the uterus of a pseudopregnant CD1 white female. Gestation in mice is 21 days. After the pups were born, mice with the highest proportion of agouti coat color were expected to have the highest proportion of ES cells in the germline.
3. Results
3.1. Homologous recombination in ES cells Out of 200 colonies, four cell lines 3D6, 4A5, 7D1, and 8A1 were found to have undergone homologous recombination. The four positive clones were identified by Southern blotting using a probe outside of the targeted region. Positive clones had a 5.0 kb Bam HI band for the knocked out allele, and a 6.5 kb band for the wild-type allele (Fig. 3). To confirm that there was no random insertion in these four positive cell lines we probed Southern blots with a fragment of the Neor gene on DNA digested with Xba I, Xho I, and Nhe I. The blots were hybridized with three different probes both inside and outside the targeted region. In addition, we used PCR to confirm that homologous recombination had taken place. The sense primer was upstream of exon 1, outside of the targeted region, and the antisense primer was in the Neor gene (data not shown).
3.2. Chromosome analysis Table 1 shows the results from chromosome analysis of the three positive ES cell lines that were injected into blastocysts. Cells with chromosome counts greater or equal to 36 were cytogenetically analyzed for the loss of the Y chromosome. All three cell lines had cells that were missing the Y chromosome: 3D6 had four out of 20 (20%), 7D1 had two (10%), and 8A1 had 16 (80%). The 7D1 cell line had one cell with what appeared to be a metaphase chromosome (centric fusion). 3D6 and
294
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
8A1 cell lines had cells with chromosome breaks and gaps. None of the cell lines showed homogeneously staining regions (HSRs), double minutes (DMs), or ring chromosomes. On the basis of the cytogenetic analysis, cell line 7D1 appeared to be the best cell line. However, this cell line yielded no chimeras. Only cell line 3D6 yielded chimeras with the ACHE knockout in the germline, a result that could not have been predicted from the chromosome analysis.
3.3. Generation of the ACHE (9) mouse From the first microinjection trial, one male chimera with 50% agouti coat color was born from the cell line 8A1. This chimera produced 100 black mice but no agouti mice. It was concluded that the ACHE knockout was not in the germline of this chimera. From the second microinjection trial, two male chimeras (3-1-I, 3-1-II) with 100% agouti coat color and one male chimera (3-1-III) with 75% agouti coat color were born from the ACHE (9 ) ES cell line 3D6. One male chimera (8-1-I) with 100% agouti coat color was obtained with the cell line 8A1.
Fig. 3. Southern blot showing one ACHE knockout allele. The 6.0 kb band is the wild-type allele. The 5.6 kb band is the knocked-out allele. The probe was outside of the targeted region
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
295
Table 1 Chromosome analysis of three ACHE ( 9) ES cell linesa Cell lines
3D6
7D1
8A1
Number of chromosomes out of 20 metaphase cells examined c35 c36 c37 c38 c39 c40 c41 c42 Modal chromosome number Number of metaphase analyzed Number of cells with (2n) modal number Percentage of cells with 2n (%) Number of cells with chromosome count \36 and minus the Y chromosome Percentage of cells \36 and minus Y chromosome (%) Percentage of metacentric chromosome (%) Presence of hypomodal cells Presence of hypermodal cells Presence of homogeneously staining regions (HSRs) Presence of double minutes (DMs) Presence of ring chromosomes Percentage of breaks and gaps (%)
0 0 0 0 8 11 1 0 40 20 11 55 4 20 0 yes yes no no no 10
0 0 0 1 0 17 2 0 40 20 17 85 2 10 5 yes yes no no no 0
1 0 0 3 13 2 1 0 39 20 2 10 16 80 0 yes yes no no no 20
a
Only cell line 3D6 produced chimeras that had the ACHE knockout allele in the germline.
All four chimeras have been bred with more than 110 C57Bl/6 females. Table 2 lists the offspring of four chimeras from the 2nd ES injection trial. To date, a total of 26 agouti mice and 73 black mice have been born. All 26 agouti coat colored mice are from chimeras 3-1-I and 3-1-III. This shows that cell line 3D6 yielded chimeras with germline transmission of the ACHE (9 ) genotype. Use of cell line 8A1 did not result in germline transmission of the ES cell genome, though it did produce chimeras. The chimera 3-1-II with 100% agouti coat color appeared to be sterile, because it has had no offspring to date. Offspring with agouti coat color have a 50% chance of carrying the ACHE knockout allele (9). Twenty-six agouti offspring were screened for ACHE (9 ) by PCR of DNA extracted from tails. Twelve out of 26 agouti mice were heterozygous ACHE knockout mice. This was determined by PCR analysis using a sense primer in the mouse ACHE gene and an antisense primer in the Neor gene. Ten heterozygous knockouts are still healthy and alive at age 29 days. Male heterozygous ACHE knockout mice and female heterozygous ACHE knockout mice will be mated to get homozygous ACHE knockout mice. From the third microinjection trial, one male chimera 3-456 with 100% agouti coat color was born from cell line 3D6. One male chimera 8-458 with 100% agouti coat color was born from cell line 8A1. Four female chimeras 8-457, 8-459, 8-460,
296
Table 2 Coat color for offspring from the mating of four male chimeras with 110 C57 Bl/6a femalesb,c C57Bl/6 female mates
Litter
Pups D.O.B
Total pups
Born dead
Born alive
Survive21 days
Black mice
Agouti mice
PCR screen agouti mice ACHE ( 9)
3-1-I 11/24/97
29
1 1 1 2
01/29/98 03/12/98 03/15/98 03/22/98
7 13 10 16
0 2 1 0
7 11 9 16
0 11 4 11
0 1 1 0
0 10 3 11
N/D 4 3 3
3-1-II 11/24/97
29
0
0
0
0
0
0
0
0
N/D
3-1-III 11/28/97
27
1 1 1 1 1 1 1
02/02/98 02/12/98 02/15/98 02/23/98 03/02/98 03/12/98 03/15/98
10 4 5 5 6 9 10
0 4 0 5 6 1 0
10 0 5 0 0 8 0
8 0 0 0 0 8 0
8 0 0 0 0 6 0
0 0 0 0 0 2 0
0 N/D N/D 2 (dead) N/D 0 N/D
8-I 12/01/97
25
1 1 1 1 2 2 2 1
02/06/98 2/12/98 2/23/98 3/5/98 3/12/98 3/20/98 3/24/98 4/1/98
8 7 7 6 20 15 10 12
0 2 2 0 4 0 0 2
8 5 5 6 16 15 10 10
8 5 0 0 16 15 10 3
8 5 0 0 16 15 10 3
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0
Total
110
23
180
29
151
99
73
26
12
a
C57 black/6 mice bought from Jackson Laboratory Inc., Bar Harbor, Maine, (207)-288-5845. N/D, not done. c The male chimeras were produced in the 2nd ES injection trial. b
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
Male, chimera D.O.B.
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
297
and 8-461 with 100% agouti coat color were born from 8A1 ACHE (9 ) ES cell line. The chromosome analysis results showed that 80% of cells in cell line 8A1 had lost the Y chromosome, which explains why this cell line produced four female chimeras. Table 3 summarizes the results from 480 blastocysts injected with three ES cell lines. Eleven chimeras were produced, of which seven were males and four were females. Two of the male chimeras transmitted the ACHE knockout in their germline.
4. Discussion Genes with a high GC content can have a low rate of homologous recombination. Therefore, it was a surprise that the mouse ACHE gene underwent homologous recombination with the high frequency of 2%. Of the 11 chimeras produced in this study, only two transmitted the ACHE knockout allele in their germline. Our finding that mice with only one ACHE allele survive to birth and adulthood is novel, because no mammalian species has previously been found to have a null ACHE allele. The wobbler mouse has 50% of normal ACHE mRNA levels [12], but the reduced level is a consequence of a mutation on chromosome 11 rather than at the ACHE locus on chromosome 5. Examples of null ACHE alleles in living animals have been reported in the Drosophila fly and the C. elegans worm. These animals survive when a single ACHE allele is deleted. However, mutations at the ACHE locus that abolish AChE activity in these animals are embryonic lethals [13,14]. The fact that mice with one null ACHE allele are healthy suggests that people may also be healthy when they have only one functional ACHE allele. There are two case reports of genetically inherited, low AChE activity in healthy families [15,16]. It is possible that these families have one deficient ACHE allele. There is no human disease of AChE deficiency, though there are two diseases where the AChE anchors are deficient, namely paroxysmal nocturnal hemoglobinuria [17] and familial myasthenia [18,19]. The absence of diseases of AChE deficiency, the importance of AChE for breathing, and the known lethality of nerve agents that inhibit AChE, have led to the expectation that the homozygous Table 3 Eleven chimeras from 480 injected blastocysts ACHE ( 9 ) cell lines
Chimeras
Males
Females
Germline transmission
3D6 4A5 7D1 8A1 Total
4 not used 0 7 11
4
0
2
0 3 7
0 4 4
0 0a 2
a
Female chimeras were not tested for germline transmission.
298
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
knockout mouse will be unable to live, and might die soon after AChE is first expressed in the mouse embryo on day 9 [20]. However, if contrary to expectation, the nullizygote does live, then it will be interesting to determine what compensatory enzymes perform the function of AChE. A prime candidate for this role is butyrylcholinesterase. During early development of the rat neural tube, butyrylcholinesterase and AChE enzymes are found in the same cells [21]. Therefore, it is possible that butyrylcholinesterase could substitute for AChE in some functions.
Acknowledgements We thank Dr. S. Steven Potter for the MJK-KO plasmid. A. Chatonnet isolated the lambda clone and prepared clones in pUC; the idea to make the knockout mouse came from Dr. Chatonnet. P. Taylor provided the mouse ACHE sequence. P. Wilder and A. Rizzino of the gene targeting facility at UNMC did the ES cell work. J. Stribley and S.H. Hinrichs of the microinjection facility at UNMC transferred the ES cells into mouse blastocysts, implanted the blastocysts, and arranged the mouse breeding. W. Xie made the gene targeting vector, screened ES colonies with Southern blots, arranged mouse breeding, and screened animals by PCR and other methods. O. Lockridge coordinated the project. Supported by U.S. Army Medical Research and Materiel Command DAMD 17-94-J-4005 and DAMD 17-97-1-7349 (to O.L.), NIH grant R01 DA011707 (to O.L.), Association Francaise Contre les Myopathies (MNM1997) (to A.C.), Nebraska State Research Initiative (to S.H.H. and A.R.) and UNMC seed grant c 98-005 (to O.L.). The opinions or assertions contained herein belong to the authors and should not be construed as the official views of the U.S. Army or the Department of Defense.
References [1] P.G. Layer, Nonclassical roles of cholinesterases in the embryonic brain and possible links to Alzheimer Disease, Alzheimer Dis. Assoc. Dis. 9 (Suppl 2) (1995) 29 – 36. [2] P.G. Layer, Non-classical actions of cholinesterases: role in cellular differentiation, tumorigenesis, and Alzheimer’s disease, Neurochem. Int. 28 (1996) 491 – 495. [3] P.G. Layer, T. Weikert, R. Alber, Cholinesterases regulate neurite growth in chick nerve cells in vitro by means of a non-enzymatic mechanism, Cell Tiss. Res. 273 (1993) 219 – 226. [4] R.T. Robertson, J. Yu, Acetylcholinesterase and neural development: new tricks for an old dog?, News Physiol. Sci. 8 (1993) 266–272. [5] W.J. Wolfgang, M.A. Forte, Expression of acetylcholinesterase during system development in Drosophila, Dev. Biol. 131 (1989) 321 – 330. [6] M.L. Hooper, Embryonic Stem Cells-Introducing Planned Changes into the Animal Germline, Harwood Academic Publishers, Chur, Switzerland, 1994, pp. 56 – 81. [7] A. Nagy, J. Rossant, R. Nagy, W. Abramow-Newerly, J.C. Roder, Derivation of completely cell culture-derived mice from early-passage embryonic stem cells, Proc. Natl. Acad. Sci. 90 (1993) 8424–8428. [8] D.M. Glover, B.D. Hames, DNA Cloning 4: Mammalian Systems, Chapter 5. Oxford University Press, New York, 1995, pp.143–184.
W. Xie et al. / Chemico-Biological Interactions 119–120 (1999) 289–299
299
[9] T.C. Hsu, K. Benirschke, An Atlas of Mammalian Chromosomes, 1, Folio 17. Springer-Verlage, New York, 1967. [10] E.J. Robertson, Teratocarcinomas and Embryonic Stem Cells — A Practical Approach, IRL Press, Oxford, 1987, pp. 108–151. [11] B. Hogan, R. Beddington, F. Costantini, E. Lacy, Manipulating the Mouse Embryo. A Laboratory Manual, Cold Spring Harbor Laboratory Press, New York, 1994, pp. 106 – 172. [12] R. Pernas-Alonso, A.E. Schaffner, C.T. Hansen, J.L. Barker, U. di Porzio, Acetylcholine esterase and peripherin mRNA level decrease in wobbler mouse, NeuroReport 6 (1995) 597 – 600. [13] J.C. Hall, D.R. Kankel, Genetics of acetylcholinesterase in Drosophila melanogaster, Genetics 83 (1976) 517–535. [14] C.D. Johnson, J.B. Rand, R.K. Herman, B.D. Stern, R.L. Russell, The acetylcholinesterase genes of C. elegans: identification of a third gene (ace-3) and mosaic mapping of a synthetic lethal phenotype, Neuron 1 (1988) 165–173. [15] R.J. Johns, Familial reduction in red-cell cholinesterase, New England J. Med. 267 (1962) 1344–1348. [16] K. Shinohara, K.R. Tanaka, Hereditary deficiency of erythrocyte acetylcholinesterase, Am. J. Hematol. 7 (1979) 313–321. [17] K. Nafa, P.J. Mason, P. Hillmen, L. Luzzatto, M. Bessler, Mutations in the PIG — A gene causing paroxysmal nocturnal hemoglobinuria are mainly of the frameshift type, Blood 96 (1995) 4650 – 4655. [18] D.O. Hutchinson, T.J. Walls, S. Nakano, S. Camp, P. Taylor, C.M. Harper, R.V. Groover, H.A. Person, D.G. Jamieson, A.G. Engel, Congenital endplate acetylcholinesterase deficiency, Brain 116 (1993) 633–653. [19] S. Camp, S. Bon, Y. Li, D.K. Getman, A.G. Engel, J. Massoulie´, P. Taylor, Patients with congenital myasthenia associated with end-plate acetylcholinesterase deficiency show normal sequence, mRNA splicing, and assembly of catalytic subunits, J. Clin. Invest. 95 (1995) 333 – 340. [20] N. Al-Fakhri, G. Bogusch, Activity of acetylcholinesterase and unspecific cholinesterase during differentiation of somites in mouse embryos, Anat. Embryol. 192 (1995) 257 – 264. [21] C. Koenigsberger, P. Hammond, S. Brimijoin, Developmental expression of acetyl- and butyrylcholinesterase in the rat: Enzyme and mRNA levels in embryonic dorsal root ganglia, Brain Res. 787 (1998) 248–258.
.