Download Genomic organization of lignin peroxidase genes of Phanerochaete

Nucleic Acids Research, Vol. 19, No. 3 599
Genomic organization of lignin peroxidase genes of
Phanerochaete chrysosporium
Jill Gaskell, Eric Dieperink+ and Daniel Cullen*
Institute for Microbial and Biochemical Technology, Forest Products Laboratory, One Gifford Pinchot
Drive, Madison, Wl 53705, USA
Received September 25, 1990; Revised and Accepted January 7, 1991
EMBL accession nos X54256 and X54257
ABSTRACT
Three lignin peroxidase (LiP) genes from the
basidiomycete Phanerochaete chrysosporium were
cloned on a single 30 kb cosmld insert. One gene,
GLG5, is the genomic equivalent of a previously
reported cDNA clone, CLG5. The other two LiP genes
are transcriptionally convergent and map to a position
approximately 15 kb downstream of GLG5. The
translational stop codons of these genes are separated
by 1.3 kb. Analysis of homokaryons established allelic
relationships to previously described LiP clones. Using
clamped homogeneous electrical field electrophoresls
(CHEF), seven chromosomal bands were resolved from
P. chrysosporium genomic DNA. On CHEF gel
Southern blots, the LiP gene family was localized to a
single, dimorphic chromosome.
INTRODUCTION
Lignin depolymerization is catalyzed by extracellular peroxidases
of white-rot basidiomycetes such as Phanerochaete chrysosporium
(1). In submerged culture, production of multiple lignin
peroxidase (LiP) isozymes is derepressed under carbon, nitrogen,
or sulfur limitation (2-4). The roles of the individual isozymes
in lignin degradation and their genetic regulation are unknown.
The lignin peroxidases of P. chrysosporium are encoded by
a family of structurally related genes (5-12), but their total
number is unclear. To date, clones from P. chrysosporium strain
BKM-1767 include the cDNA encoding isozyme H8 (5), the
corresponding genomic clone (9), cDNA's designated CLCJ4 and
CLG5 (6), genomic clone GLG2 which is nearly identical to
CLG5 (Y. Zhang, 1987, Ph.D. Thesis, Michigan State
University), genomic clones V4 and O282 (11), genomic and
cDNA clones designated ML4 (10), and most recently, genomic
clone LPOB which is closely linked to H8 ( = LPOA). A large
family of structurally similiar genes has also been identified in
P. chrysosporium strain ME446 (8). An allelic relationship
between H8 and ML4 has been proposed (10) but direct
experimental evidence is lacking. The genomic organization of
LiP genes is also poorly understood. This is in part due to the
lack of a well-defined recombination system in P.
chrysosporium^ 13).
Wild-type P. chrysosporium strains are heterokaryotic, and
following meiosis haploid recombinants can be recovered as
homokaryotic basidiospores (13 — 15). Recombinants vary
substantially with respect to numerous phenotypic traits, including
pigmentation, sporulation, and lignin peroxidase activity (16,
unpublished). Analyses of single basidiospore cultures have been
used to exclude allelic relationships among several LiP clones
(11) and to generate a genetic map using restriction fragment
length polymorphisms (RFLP's)(16). The latter study involved
strain ME446 and utilized a 33-mer oligonucleotide which
hybridized to eight different genomic LiP clones under stringent
conditions.
We report here the genomic organization of three very closely
linked lignin peroxidase genes. In order to investigate the
chromosomal organization of the entire LiP gene family, P.
chrysosporium chromosomes were resolved using pulse field
eletrophoresis and the LiP genes were localized to a single,
dimorphic chromosome.
MATERIALS AND METHODS
DNA cloning and characterization
A 6,000-member library of P. chrysosporium strain BKM-1767
(ATCC24725) was prepared in cosmid pWE15 (17). Clones were
maintained at -90°C in 96-well microtitre plates (18). The
library was probed with a nick-translated genomic clone encoding
LiP isozyme H8 (11) at low stringency (35% formamide, 37°Q.
Standard Southern and colony hybridization techniques were used
throughout (19-21).
Cosmid restriction sites were mapped directly as described by
Wahl et al. (17) and by analyses of fragments subcloned into
Bluescript n vectors (Stratagene, La Jolla, CA). Double stranded
templates were sequenced by the dideoxy chain termination
method (22) using universal primers or synthetic oligonucleotides
synthesized by the University of Wisconsin Biotechnology Center.
Nucleotide and amino acid alignments were generated by the
method of Wilbur and Lipman (23) using DNASTAR Inc.
software (Madison, WT).
To ascertain allelic relationships, BssHU. site polymorphisms
(Fig.l) were exploited by Southern hybridization of
homokaryons. Strain BKM-1767 was fruited as described (14)
* To whom correspondence should be addressed
+
Present address: University of Minnesota Medical School, Minneapolis, MN, USA
600 Nucleic Acids Research, Vol. 19, No. 3
and single, germinating basidiospores were isolated using a
dissecting microscope. DNA was extracted from parent and single
basidiospore cultures (24), digested with BssHU, size fractionated
in 0.8% agarose, blotted to Nytran (S&S, Keene, NH), and UV
crosslinked (UV Stratalinker 1800, Stratagene Inc.). BssHU
digests of cosmid pQQ24 (Fig. 1, top map) and of a plasmid
containing the LPOB/LPOA genes (Fig. 1, bottom) were included
as size standards.
The blot was probed twice; first with a nick-translated 1081
bp BglWKpnl fragment derived from pQQ24 and then with a
28-mer oligonucleotide (5'-AAACCTTTTGATTCGCGATTTTTAATTG-3'). The fragment included 516 bp LiPA
coding region plus 565 bp 3' untranslated intergenic region (Fig.
1). Under the hybridization stringencies employed, this probe
would differentiate LiPA (2353 bp) from its' allele LPOA (2573
bp), and it would also hybridize to LiPB (1475 bp) and its' allele
LPOB (1593 bp). The 28-mer hybridizes to a position 600 bp
3' of LiPA's translational stop codon. This region of intergenic
LiPB-LiPA shows no homology to LPOB-LPOA (see below).
The 28-mer was synthesized by the University of Wisconsin
Biotechnology Center and further purified by acrylamide gel
electrophoresis and reverse phase chromatography (21).
Chromosomal location of LiP genes
Clamped homogeneous electric field (CHEF) electrophoresis was
performed on a BIO RAD CHEFII apparatus as described by
Chu et al (25). Plug preparations, run conditions, and blotting
were similiar to Brody and Carbon (26) with the following
modifications. To prepare protoplasts, 1 x 10s conidiospores
were germinated in modified Vogel's medium, washed by
centrifugation, digested for 3 hr at 37°C in 0.7 M MgSO4, 10
mg/ml Novozyme 234 (Novo Labs), 50 mM maleic acid pH 5.9,
and filtered through Miracloth (Calbiochem Inc). Protoplasts in
the filtrate were washed three times with 1.2 M sorbitol by gentle
centrifugations ( < l,000xg). The final pellet was resuspended
in GMB buffer (0.125 M EDTA pH 7.5, 0.9 M sorbitol) such
that the final protoplast concentration was l x l O 8 cells/ml.
Protoplasts were equilibrated to 37°C, mixed with an equal
volume of Beckman LMP agarose (1.4%) in GMB, and then
poured into a 1X9X20 mm plug mold. Protoplasts were lysed
in plugs and stored as described (26).
Gels were prepared using 0.8% SeaKem GTG (FMC) agarose
in 0.5 X TAE. Final gel dimensions were 14 X13 X1 cm. Plugs
B
B
K
were trimmed to approximately 1 x 9 x 10 mm and then sealed
into wells (1.5x10x10 mm) with 0.8% SeaKem GTG. Run
conditions were as described (26) except pulsing intervals and
their duration were: 50 min/80 hr; 45 min/13.5 hr; 37 min/80 hr.
Gels were stained with ethidium bromide, destained, blotted to
Nytran, and UV crosslinked.
Nick-translated probes included cosmids pQQ24 (Figure 1),
pO282 (11), pFFF62, which contains LiP gene V4 (11), and a
plamid containing the LPOA ( = H8) gene (9,12). Specific
activities of probes exceeded l x l O 8 dpm//tg. Hybridizations
were at 50°C in 50% formamide, 7% SDS, 0.125 M Na2HPO4,
1 mM EDTA. Final washes were at 60°C in 1 % SDS, 0.025
M Na2HPO4.
A 33-mer (M33) that hybridizes to at least 8 LiP clones of
P. chrysosporium strain ME446 (8,27) was also used. It was
synthesized, purified, end-labeled, separated from unincorporated
label,and hybridized as descibed above except that 1 X107
cpm/ml hybridization buffer were used. Final washes were at
60°C in IX SSC, 0.1% SDS.
RESULTS
Isolation and characterization of closely-linked LiP genes
When the cosmid library was probed under low stringency with
the gene encoding LiP isozyme H8 ( = LPOA), approximately
30 clones hybridized with varying intensities. Subsequent
restriction digests and Southern analyses of these cosmids showed
that many contained previously described genes (5,6,10) and that
several contained more than one H8-homologous region. One
clone, designated pQQ24, gave a particularly strong signal even
after extensive washing and was found to contain 3 intact LiPlike regions. This cosmid was mapped in detail, subcloned, and
A
2 3 4 5 6
2573X
2353-
1 2 3 4 5 6
2353-
1475
B
UPB
UPA
LPOB
LPOA
Figure 1. Partial restriction map of aJlelic LiP gene clusters. Top map shows
approximately 8 kb from cosmid pQQ24 which contains 3 intact LiP-like regions.
The distance between GLG5 and LiPB is approximately 14 kb. The allelic
homologue is shown in the bottom map (12). No restriction sites have been mapped
5' to LPOB. LPOA is identical to H8 (9). Abbreviations: B, BssHU; K, Kpnl;
R. EcoRl.
Figure. 2. Southern blot of DNA from BKM and single basidiospore cultures.
In Panel A the blot was probed with the nick-translated KpnllBglU fragment of
LiPA (Fig. I). Hybridization was at 50°C in 50% formamide, 7%SDS, 0.25M
Na2HPO4, lmM EDTA. Final washes were at SOX in 2% SDS in 0.25M
Na2HPO4. In Panel B, the blot has been stripped of label and reprobed with the
LiPA-specific 28-mer. After end-labeling with T4 polynucleotide kinase (21),
unincorporated label was removed by exclusion chromatography (Push Columns;
Stratagene Inc). The specific activity exceeded 1 x 10s dpm//ig. Hybridization
conditions for the oligonucleotide included prehybridization in 6 x SSC, 0.1%
SDS, 5x Denhardts, 0.01% denatured salmon sperm DNA for 2 hr at 42°C
followed by hybridization in 6 x SSPE, 0.1% SDS, 5 x Denhardts at 42°C with
1 x 106 dpm/rnl. Lane 3 contains HindUl digested Lambda DNA. All other lanes
contain BssHll digestions: lanel, plasmid containing LPOB-LPOA genes; lane
2, cosmid pQQ24; lane 4, dikaryon BKM-1767; lane 5, single basidiospore strain
No. 2; lane 6, single basidiospore No. 10. Both blots were exposed to Kodak
XAR-5 film with amplifying screen overnight. Sizes of BssHU fragments are
shown in left margins in base pairs.
Nucleic Acids Research, Vol. 19, No. 3 601
approximately 8 kb were sequenced (Figure 1). Southern blots
of genomic DNA digested with Clal plus Notl showed the 26
kb LiP-containing region to be contiguous (data not shown).
Excluding the intron sequences, the nucleotide sequence of
GLG5 (EMBL accession No. X54256) is identical to a previously
published cDNA clone, CLG5 (6). GLG5 is remarkably similiar
to another genomic clone, GLG2 (Y. Zhang, 1987, Ph.D. Thesis,
Michigan State University). The nucleotide and amino acid
alignments of these two clones are 97.7% and 100% identical,
respectively. All nucleotide mismatches are located in the third
postions of codons or within introns. Both GLG2 and GLG5
contain nine introns as opposed to eight found in other LiP clones
(7-12). GLG2 and GLG5 may be allelic forms of the same gene.
Two transcriptionally convergent LiP-like clones, LiPA and
LiPB, are located approximately 14 kb downstream of GLG5
(Figure 1). The nucleotide sequence of the entire region was
determined (EMBL ascession No. X54256). This region is
strikingly similiar in structure to a recently reported XEMBL3
subclone which contains two lignin peroxidase genes designated
LPOA and LPOB (Figure 1)(12). The overall nucleotide sequence
identity between the LPO and the LiP clones, including the
intergenic region, is 93%. The translational stop codons of the
LiP and LPO genes are separated by 1360 bp and 1322 bp,
respectively. Within this region the nucleotide identity is 87%.
Intron numbers and their relative positions are conserved between
the homologous LPO and LiP genes.
Sequence comparisons reveal LiPA to be identical to ML-4
(10) and LPOA identical to LiPH8 (9,10). The nucleotide
An
sequence identity between the LiPA and LPOA coding regions
is 99%. Their deduced amino acid sequences are identical except
for a single amino acid change in the signal sequence. In contrast,
LiPB and LPOB nucleotide and amino acid sequences are 96.5 %
and >99% identical, respectively. The single amino acid
difference is at position 298 with a Leu residue in LiPB and a
His in LPOB.
The close structural similiarities between the LPO and LiP
clones suggested an allelic relationship. Segregation of BssHQ
bands is consistent with allelism (Figure 2A). Bands specific to
LIPB (2353 bp), LiPA (1475 bp), LPOB (2573 bp) and LPOA
(1593 bp) are evident in the dikaryon Qane 4) and then segrgate
into the homokaryotic cultures 0anes 5 and 6). Higher molecular
weight bands visible in lanes 4—6 represent related LiP genes
such as 0282 (11). Segregation of allelic variants was confirmed
HS
I
O282
pFFF02
pOQ24
H33
••*
Figure. 4. Chromosomal location of LiP genes in P. chrysosporium BKM-1767.
Arrows to the left of the ethidium bromide stained gel show the position of
hybridizing bands at approximately 3.5 and 3.7 mb. Nytran Wots were probed
with the indicated genes or 33-mer as described in Materials and Methods. Kodak
XAR film (right) was exposed with amplifying screens for 24 to 72 hr.
B
2
Figure. 3. CHEF chromosome separations. Gel A, P. chrysosporium (Pc), and
standards S. pombe (Sp), S. cerevisiae (Sc). Gel B, A. nidulans (An) run as
standard. Chromosomal plugs prepared and run as described in Materials and
Methods. Estimated size of chromosome standards in mb shown in right margins.
3
4
Figure. 5. Segregation of homologous chromosomes in single basidiospore
cultures. Ethidium bromide stained gel (Panel A) and Nytran blot probed with
nick-translated pQQ24 (Panel B) as described in Materials and Methods. Lanel
contains A. nidulans as molecular weight standard. Lanes 2,3, and 4 contain DNA
from dikaryotic BKM-1767, single basidiospore culture No. 2 and No. 10,
respectively. Arrows in lane 2 indicate position of the 3.5 and 3.7 mb hybridizing
bands. Kodak XAR film was exposed with amplifying screen for 2 days.
602 Nucleic Acids Research, Vol. 19, No. 3
by probing the blot with the 28-mer (Fig. 2, panel B). The cosmid
standard (lane 2), dikaryon (lane 4), and only one of the
homokaryons(lane 6) hybridize.
CHEF gel analysis
Under the CHEF conditions described above, at least seven
chromosome bands were resolved from P. chrysosporium DNA
(Figure 3). At least two of these are doublets, as judged by the
relative intensity of ethidium bromide staining. Based upon
comparisons with Saccharomyces cerevisiae, Schizosaccharomyces pombe, and Aspergillus nidulans, the P. chrysosporium
bands are approximately 2.0, 2.8, 3.2, 3.6, 3.8, 4.4, 4.8 mb.
The band pattern is similiar for strain ME446 (dat not shown),
but differs substantially among homokaryons because of meiotic
recombination and chromosome polymorphisms. For example,
basidiospore culture No. 2 (Fig. 5, lane 3) has bands at
approximately 4.3, 3.7, 3.5, 3.0, 2.8, and 2.0 mb. Assuming
doublets at 2.8, 3.0, and 3.7 mb the total genome size is
approximately 29 mb, somewhat lower than the previous estimate
of 36.5 mb for the haploid genome<27).
Six different LiP clones from BKM-1767 hybridized to
electrophoretic bands at 3.7 and 3.5 mb (Figure 4). The clones
were H8 (=LPOA), O282, V4 (pFFF62), and the three genes
residing on pQQ24 (Fig. 1). In addition, the 33-mer (M33) which
hybridizes to 8 of the 11 LiP clones of ME446, hybridizes to
the same two bands. These results could be interpreted as
interchromosomal duplications/translocations of LiP genes.
Alternatively, homologous chromosomes in the dikaryon might
be dimorphic with respect to electrophoretic mobility. The latter
explanation is supported by the presence of single bands in
homokaryotic derivatives (Figure 5). This, in turn, is enirely
consistent with the identification of specific alleles in these
cultures (Fig. 3). Subsequent probing of the same CHEF gel
with genes encoding glyoxal oxidase and a cellobiohydrolase have
identified other dimorphic chromosomes (P. Kersten and S.
Covert, unpublished data). Chromosome length polymorphisms
have been observed in Plasmodiumjalciparum (28), S. cerevisiae
(29), Candida albicans (30),Ustilago maydis (31), and Coprinus
cinerus (M. Zolan, personal communication). The latter two fungi
belong to the same taxonomic class as P. chrysosporium, the
Basidiomycetes.
No evidence for a second LiP cluster was observed using the
six LiP clones as probes. In addition, the 33-mer that hybridizes
to two RFLP-defined clusters in ME446 (27), hybridized to the
3.7 and 3.5 mb bands of BKM-1767 (Figure 4) and ME446 (data
not shown).
DISCUSSION
The cosmid and CHEF analyses described here clearly
demonstrate clustering of LiP genes at both kilobase and
megabase levels. The clustering of LiP genes is consistent with
the RFLP finding that large variations in LiP activity among
haploid recombinants do not correlate with the distribution of
LiP alleles (16). Recombination between closely linked structural
genes is expected to be relatively low and the cluster would
segregate as a unit. (Accurate estimates of recombination
frequencies would require complete mapping of all LiP genes
on the chromosome.) Variation among strains is undoubtedly due
to complex interactions involving many genes, linked and
unlinked. Raeder et al. (16) detected one locus unlinked to LiP
genes that exerted a significant effect on LiP activity.
Other structural genes probably involved in lignin
depolymerization include those encoding Mn-dependent
peroxidases (32,33), glyoxal oxidase (34), and perhaps additional
LiP genes. It is unlikely that unidentified LiP genes are present
outside the dimorphic chromosome, unless their nucleotide
sequences are substantially different from the six clones and
33-mer used to probe CHEF gels in this study.
The CHEF separation described here will allow rapid and
convenient chromosome mapping of a variety of markers in wild
type and genetically transformed (35) strains. For example, we
have mapped the genes encoding glyoxal oxidase and several
cellulase genes to chromosomes other than those carrying the
LiP cluster (P. Kersten and S. Covert, unpublished). Some of
the cellulase genes are very closely linked; i.e. within 1 kb
(Covert and Cullen, submitted).
The mechanism(s) giving rise to chromosome length
polymorphisms may also be related to variations in LiP activity.
In P. falicurum, polymorphisms are caused by homologous
recombination in distal regions of chromosomes, and it has been
suggested that recombination in these
dynamic regions might be related to antigenic variation (28).
Extensive mapping of the chromosome homologues will be
required to determine if a similar mechanism is operative in P.
chrysosporium. These investigations will be greatly facilitated
by the preparation of chromosome-specific genomic libraries from
haploid recombinants, now possible using the CHEF separation
we describe together with other recent techniques (e.g., 36,37).
ACKNOWLEDGEMENTS
The authors thank P. Kersten and S. Covert for their thoughtful
comments and M. Zolan for helpful advice and generously
supplying unpublished procedures. This work was supported in
part by U.S. Department of Energy Grant DE-FGO2-87ER13712
to D. Cullen and T. K. Kirk.
REFERENCES
1. Kirk, T. K. and Farrell, R. L. (1987) Ann. Rev. Micrbiol., 41:465-505.
2. Kirk, T. K., Farrell, R. L., Croan, S. and Tien, M. (1986) Enz. Microb.
Techno!., 8:27-32.
3. Leisoia, M. D., Kozulic, D., Meusdierffer, D. and Fiechter, A. (1987) J.
Bid. Chem., 262:419-424.
4. Jeffries, T. W., Choi, S. and Kirk, T. K. (1981) Appl. Environ. Microbiol.,
42:290-296.
5. Tien, M. and Tu, C. P. (1987) Nature 326:520-523.
6. deBoer, H. A., Zhang, Y. Z., Collins, S. and Reddy, C.A. (1987) Gene
60:93-102.
7. Asada, Y., Kimura, Y., Kuwahara, M., Tsukamoto, A., Koide, K., Oka,
A. and Takanomi, M. (1988) Appl. Microbiol. Biotechnol., 29:469-473.
8. Brown, A., Sims, P., Raeder, U. and Broda, P. (1988) Gene 73:77-85.
9. Smith, T., Schalch, H., Gaskell, J., Covert, S. F. and Cullen, D. (1988)
Nucleic Acid Res., 16:1219.
10. Andrawis, A., Pease, E., Kuan, I., Holzbaur, E. and Tien, M. (1989)
Biochem. Biophys. Res. Comm. 162:673-680.
11. Schalch, H., Gaskell, J., Smith, T. and Cullen, D. (1989) Mol. Cell. Biol.,
19:2743-2747.
12. Huoponen, K., OUikka, P., Kalin, M., Walther, I., Mantsala, P. and Reiser,
J. (1990) Gene 89:145-150.
13. Alic, M. and Gold, M. H. (1985) Appl. Environ. Microbiol., 50:27-30.
14. Gold, M. H. and Cheng, T. M. (1979) Arch. Microbiol., 121:37-41.
15. Thompson, W. and Broda, P. (1987) Trans. Br. Mycol. Soc. 3:285-294.
16. Raeder, U., Thompson, W. and Broda, P. (1989) Mol. Microbiol.,
3:919-924.
17. Wahl, G. M., Lewis, K. A. Ruiz, J. C , Rothenberg, B., Zhao, J. and Evans,
G. A. (1987) Proc. Nat. Acad. Sci. U.S.A., 84:2160-2164.
18. Vollmer, S. J. and Yanofsky, C. (1986) Proc. Nat. Acad. Sci. U.S.A., 83:
4867-4873.
Nucleic Acids Research, Vol. 19, No. 3 603
19. Church, G. M. and Gilbert, W. (1984) Proc. Nat. Acad. Sci. U.S.A.,
81:1991-1995.
20. Taub, F. and Thompson, E. B. (1982) Anal. Biochem., 126:222-230.
21. Sambrook, J., Fritsch, E. F. and Manaitis, T. (1989) Molecular Cloning:
A Laboratory Manual. Cold Spring Harbor Press, Cold Spring Harbor.
22. Sanger, F., Nicklen, S. and Coulson, A. R. (1977) Proc. Nat. Acad. Sci.
U.S.A., 74:5463-5467.
23. Wilbur, W. J. and Lipman, D. J. (1983) Proc. Nat. Acad. Sci. U.S.A.
80:726-730.
24. Specht, C , DiRusso, C , Novotny, C. and Ulrich, R. (1982) Anal. Biochem.
119:158-163.
25. Chu, G., Vollrath, D. and Davis, R. W. (1986) Science 234:1582-1585.
26. Brody, H. and Carbon, J. (1989) Proc. Nat. Acad. Sci. U.S.A.
86:6260-6263.
27. Raeder, U., Thompson, W. and Broda, P. (1989) Molecular Microbiology
3:911-918.
28. Corcoran, L. M., Thompson, J. K., Walliker, D. and Kemp, D. J. (1988)
Cell 53:807-813.
29. Ono, B. and Ishino-Arao, Y. (1988) Curr. Genet., 14:413-418.
30. Rustehenko-Bulgac, E.P., Sherma, F. and Hicks, J. B. (1990) J. Bact.
172:1276-1283.
31. Kinscherf, T. G. and Leong, S. A. (1988) Chromosoma 96:427-433.
32. Pribnow, D., Mayfield, M. B., Nippe, V. J., Brown, J. A. and Gold, M.
H. (1989) J. Biol. Chem., 264:5036-5040.
33. Pease, E. A., Andrawis, A. and Tien, M. (1989) J. Biol. Chem.,
264:13531-13535.
34. Kersten, P. (1990) Proc. Nat. Acad. Sci. U.S.A., 87:2936-2940.
35. Alic, M., Clark, E. K., Kornegay, J. R. and Gold, M. H. (1990) Curr.
Genet., 17:305-311.
36. Anand, R., Villasante, A. and Tyler-Smith, C. (1989) Nucl. Acids Res.,
17:3425-3433.
37. van de Pol, T.J.R., Cremers, F.P.M., Brohet, R.M., Wieringa, B. and
Rapers, H.H. (1990) Nucl. Acids Res., 18:725-731.