African Journal of Biotechnology Vol. 10(84), pp. 19590-19597, 26 December, 2011 Available online at http://www.academicjournals.org/AJB DOI: 10.5897/AJB10.2464 ISSN 1684–5315 © 2011 Academic Journals
Full Length Research Paper
Strain improvement in Trichoderma viride through mutation for overexpression of cellulase and characterization of mutants using random amplified polymorphic DNA (RAPD) Shazia Shafique*, Rukhsana Bajwa and Sobiya Shafique Institute of Plant Pathology, University of the Punjab, Quaid-e-Azam Campus, Lahore, Pakistan. Accepted 30 September, 2011
Cellulases are multi-enzymatic complex proteins that catalyze the conversion of cellulose to glucose. Indigenous strain of Trichoderma viride FCBP-142 was selected to develop over-producer of cellulases and subjected to mutagenesis with ultra violet (UV) and chemical ethyl methane sulfonate (EMS). Among 178 survivals after UV irradiation, bigger zones of clearing on agar plates appeared around 81 colonies of putative mutant strains of native fungus with maximum of 87 IU/ml by Tv-UV-5.6 strain in comparison to parental strain (53 IU/ml). For EMS treatment, the enzyme production by the most active mutant Tv-Ch-4.3 showed even higher-level of cellulase activity (122.66 IU/ml) in contrast to the UV and parental strains. Genetic relationships of stable mutants of T. viride were also analyzed with RAPDPCR. Results obtained from the comparison between genotypes of T. viride exhibited differences in sizes and numbers of amplified fragments per primer for each isolate. The dendrogram showed that the genotypes Tv-Ch-4.3 and Tv-Ch-5.5 were distinctly classified into one category, while the two isolates (Tv-UV-5.6 and Tv-UV-5.9) of T. viride FCBP-142 (parental) have the nearest genetic relationship. Moreover, the five isolates from T. viride genotypes shared an average of 75 percent bands. Key words: Cellulase activity, Trichoderma viride, UV irradiation, EMS treatment, RAPD-PCR. INTRODUCTION Cellulases are employed in industries for the preparation of medicines, resins, perfumes, starch production, waste treatment and baking etc. (Sun and Cheng, 2002; Beauchemin et al., 2003). Considering the industrial potentials of cellulases, an important aspect of cellulose research is to obtain highly active cellulases at low cost. Therefore, the primary goal of research remains reducing the cost of cellulases acting on pretreated biomass used for bio-ethanol production by increasing the specific activity of the enzymes in the cellulase complex. Biotechnological companies are now searching for modified properties of proteins or enzymes to enhance their efficiency because of their multipurpose applications, and some notable success has already
*Corresponding author. E-mail:
[email protected].
been achieved by protein/enzyme engineering (Schulein, 2002). Moreover, in spite of huge extent of research for finding more active enzyme preparations from a large variety of microorganisms, the enzymatic saccharification of cellulose so far has not been reached to the level of conversion of cellulose to glucose by the microbial enzymes. Therefore, several approaches including chemical mutation, UV irradiation and genetic engineering to obtain enhanced cellulase producing strains have been given a high priority in the last decade (Kotchoni and Shonukan, 2002). Industrial application success for improved strains depends on their genetics and physiological characterization with a system that allows quick diagnosis. These diagnostic procedures for mutation are based upon a number of techniques in which resistant type of mutants can also be used for this purpose.
Shafique et al.
The random amplified polymorphic DNA (RAPD) procedure introduced by Williams et al. (1990) is technically the simplest version. It is considered a sensitive and quick technique to distinguish the differences between closely related species and to analyze differences between mutant strains of the same species (Jones and Kortenkamp, 2000). The present work was therefore undertaken to develop stable mutant strains of Trichoderma viride FCBP-142 for the production of cellulases following UV and chemical mutagenesis and to resolve the genetic variability of mutant derivatives with their wild strain using RAPD markers.
19591
fermentation. Enzyme assay Reducing sugars in culture broth were determined by 33, 5dinitrosalicylic acid (DNS) method of Miller (1959) with glucose as standard. The reaction was allowed to proceed by incubating 0.5 ml of enzyme solution with 0.5 ml of 1% substrate in 0.05 M citrate buffer at pH 4. The reaction mixture was incubated on a shaking water bath for 15 min at 50°C. Sugar content in the reaction mixture was measured with DNS and absorbance was taken at 600 nm on a spectrophotometer using a calibration curve obtained with glucose. Enzymatic activity of cellulase complex was expressed as International Units/ ml, which is defined as the amount of enzyme that releases one micro mole of reducing sugar expressed as glucose per minute.
MATERIALS AND METHODS The strains of T. viride (FCBP-142, FCBP-167 and FCBP-232), T. reesei (FCBP-84, FCBP-271, FCBP-364) and T. harzianum (FCBP125, FCBP-139, FCBP-140, FCBP-193, FCBP-210, FCBP-325) were obtained from stock cultures of First Fungal Culture Bank of Pakistan, Department of Mycology and Plant Pathology, University of the Punjab, Lahore. These cultures were screened and studied for cellulase enzyme activity using solid states as well as submerged fermentation techniques using Mendel’s mineral salt solution that is: Urea 0.3 g/l, (NH4)2SO4 1.4 g/l, KH2 PO4 2.0 g/l, CaCl2 0.3 g/l, MgSO4 0.3 g/l, yeast extract 0.25 g/l and proteose peptone 0.75 g/l with 10 g/l of cellulose (Shafique et al., 2009). Among the strains, potential strain was selected for further study.
Genetic characterization Two of the most efficient cellulolytic strains from UV mutants and two from chemical mutants were selected. Investigative studies were performed to elucidate the relationships of differences between mutant strains and wild type (parental) by RAPD-PCR. DNA isolation The modified CTAB (Cetyltrimethylammonium bromide) method of Saghai-Maroof et al. (1984) was used for DNA extraction. The quality and yield of the extracted DNA was assessed via spectrophotometry and gel electrophoresis.
Induction of mutation Improvement in enzyme production by mutagenesis was sought to isolate hyper-producer mutant derivatives of T. viride FCBP-142. For this, the selected fungal strain of T. viride FCBP-142 was treated with mutagenic agents - ultraviolet (UV) irradiation and ethyl methanesulfonate (EMS) by using following protocols.
UV mutagenesis Ten milliliters of conidial suspension (5 × 105 conidia/ml) from a week-old potato dextrose agar (PDA) plates was transferred to the sterilized Petri plates and exposed to ultraviolet irradiation for 40 min with 5 min time interval under UV lamp having a wavelength of λ = 254 nm and 220 V at 50 Hz. The distance between lamp and suspension was adjusted to 20 cm for each trial (Hamad et al., 2001). After the time intervals, 200 µL of the conidial suspension was transferred to the Petri plates containing PDA with the addition of 0.1% Triton X-100 and L-sorbose as colony restrictors. Plates were then incubated at 30 ± 2°C and Isolated colonies were replicated on to the plate-screening medium.
Chemical mutagenesis Chemical mutagenesis was carried out according to the method of Morikawa et al. (1985). Appropriate dilutions of 50, 100, 150, 200 and 250 were prepared from 300 µg ml -1 of EMS solution, and then 5 ml of adjusted conidial suspension was treated with each concentration of EMS solution for 30 min at 37°C in a water bath shaker. Two hundred microliters of mutagenized sample was plated to obtain survivors. Finally, the screened out mutant derivatives were then assayed quantitatively for enzyme activity by shake flask
RAPD-PCR This protocol was modified after Williams et al. (1990). PCR amplification conditions were optimized in a GeneAmp-2700 thermocycler (Applied Biosystems, 850 Lincoln Centre Drive, Foster City, C. A., USA). A standard RAPD-PCR reaction was performed in volume of 25.0 µL, containing 50 mM KCl, 10 mM Tris-HCl pH 8.3, 1.5 mM MgCl2, 0.01% (w/v) gelatin, 2 mM of each dNTP (Fermentas Inc. 7520 Connelley Drive, Maryland 21076, USA), 40 pmol of each primer (BioBasic, USA), 1.5 U of Taq DNA polymerase (Fermentas, USA) and 30 ng of DNA template. PCR profile was programmed at 94°C for 1 min denaturation followed by primer annealing at 35°C for 1 min and primer extension at 72°C for 2 min with a total of 40 cycles. The initial denaturation of DNA was for 2 min at 94°C. The final extension period was adjusted for 5 min at 72°C, then 3 µL of 6x loading buffer was added to each tube. These PCR products were analyzed on 1% agarose gel.
Data analysis Treatments effects of UV irradiation and chemical doses were compared and subjected to one way analysis of variance followed by LSD method, to demarcate mean differences (Steel and Torrie, 1980). A name or a number was designated for each RAPD marker based on the molecular size and primer used. Intensely stained DNA bands on agarose gel ranging from 0.2 to 3.0 Kb were recorded. The presence or absence of a DNA fragment was scored using a binary system of 0 (in the absence of the band) and 1 (if the band is present). Similarities among fungal genotypes were deduced on the basis of the number of shared amplification products (Nei and Li, 1979). The data for genetic variability was analyzed by applying Minitab (2004) computer software program.
19592
Afr. J. Biotechnol.
Table 1. Screening of Trichoderma viride FCBP-142 mutants for cellulase activity isolated after different time intervals of UV irradiation.
S/N 1 2 3 4 5 6 7 8 9
Exposure (min) 0 5 10 15 20 25 30 35 40
Number of survivor All 51 44 30 22 17 10 4 Nil
Number of improved mutant 0 15 13 12 11 9 2 Nil Nil
Range of cellulase activity (IU/ml) 53.00 ± 0.1588 52.00 – 59.41 ± 0.1560 54.00 – 64.26 ± 0.1619 59.00 – 67.00 ± 0.1179 66.00 – 76.40 ± 0.1947 74.35 – 87.00 ± 0.1531 71.50 – 77.00 ± 0.0882 71.70 – 73.46 ± 0.0900 Nil
Each value is an average of three replicates; ± indicated the standard error from mean value.
Table 2. Screening of Trichoderma viride FCBP-142 mutants for cellulase activity isolated after different doses of EMS.
S/N 1 2 3 4 5 6 7
EMS concentrations (µg/ml) 0 50 100 150 200 250 300
Number of survivor
Number of improved mutant
All 87 68 39 27 14 3
0 23 16 19 14 11 1
Range of cellulase activity (IU/ml) 53.00 ± 0.1588 60.00 – 75.00 ± 0.1179 73.24 – 81.60 ± 0.1179 79.24 – 94.62 ± 0.1334 99.00 – 122.66 ± 0.1172 117.00 – 121.77 ± 0.0867 98.39 ± 0.1531
Each value is an average of three replicates; ± indicated the standard error from mean value.
RESULTS For the enhancement of cellulolytic activity by the fungus, conidial suspension was subjected to UV irradiations. Among the 178 survivals, 81 mutants of the native fungus T. viride FCBP-142 were selected for hyper-production of extra cellular cellulases. The selection of these cellulaseproducing strains was based on the larger diameter of clear zone surrounding the colonies on plate screening medium as compared to wild strain. This screening step was found to give fairly reliable indication of exhibited cellulolytic activities. The efficient mutants were further assessed in shake flask cultures. The strain designated as Tv-UV-5.6 exhibited maximum cellulase activity of 87 IU/ml at an exposure time of 25 min of UV irradiation. Further increase in UV exposure time caused a decline in activity, while the highest increase in UV treatment up to 40 min however resulted in the death of fungus (Table 1). Furthermore, improvement in enzyme production by EMS mutagenesis was carried out to isolate hyperproducer mutant derivatives of T. viride FCBP-142 using different doses of EMS (50-300 µg/ml) for 30 min and pertaining results are presented in Table 2. After the chemical mutagenesis, 84 mutant derivatives were isolated out of 238 survivals, for their ability to hydrolyze
the cellulose on agar plates more efficiently and significantly in comparison to parental. These selected mutants were subjected to quantitative analysis by shake flask cultures. The 200 µg/ ml of EMS concentration evidenced the most promising concentration as it illustrated significantly higher cellulase activity (122.66 IU/ ml) than the wild type (52.97 IU/ml) by mutant strain TvCh-4.3 (Table 2). Data records demonstrated the exponential increase in cellulolytic activities by UV and EMS mutant derivatives as compared to wild type strain at different time intervals (Figure 1). However, the trend in increase in cellulase activity of mutants was found similar to wild type strain and attained its peak after 72 hrs of incubation period. More also, the top two mutants from UV treatment as well as from EMS treatment were selected and RAPDPCR analysis of genomic DNA was performed to detect genetic diversity of these mutants with the wild type by using 20 random primers. The primer-wise details of DNA polymorphism detected in T. viride FCBP-142 genotypes are elaborated in Table 3. It was clearly depicted from the results that quality and quantity of amplification products were sufficient for detection of genetic distance among the T. viride genotypes. Each of the primers produced distinct polymorphic banding patterns in all the genotypes
19593
U/ml
Shafique et al.
Figure 1. Comparison between parental strain of Trichoderma viride FCBP-142 and its UV and EMS treated mutants for cellulase activity.
Table 3. DNA polymorphism detected in mutant genotypes of Trichoderma viride FCBP-142.
RAPD primer M-01 M-02 M-03 M-04 M-05 M-06 M-07 M-08 M-09 M-10 M-11 M-12 M-13 M-14 M-15 M-16 M-17 M-18 M-19 M-20
Sequence (5′′ to 3′′) GTTGGTGGCT ACAACGCCTC GGGGGATGAG GGCGGTTGTC GGGAACGTGT CTGGGCAACT CCGTGACTCA TCTGTTCCCC GTCTTGCGGA TCTGGCGCAC GTCCACTGTG GGGACGTTGG GGTGGTCAAG AGGGTCGTTC GACCTACCAC GTAACCAGCC TCAGTCCGGG CACCATCCGT CCTTCAGGCA AGGTCTTGGG
Monomorphic DNA fragment 35 15 15 5 10 15 25 5 10 5 10 15 10 10 15 5 5 5 5 15
examined. The level of polymorphism was different with each primer among the genotypes (Figure 2). It is also evident from Figure 2A that band 1 of 1600 bp in wild strain had been inactivated in all the mutants, while
Polymorphic DNA fragment 14 12 17 13 29 20 7 13 16 14 9 18 15 10 10 9 10 17 12 23
Size (bp) 250 - 1500 300 - 1900 350 - 1600 350 - 2050 200 - 2500 300 - 1750 400 - 2750 300 - 1100 300 - 2000 450 - 1750 400 - 1800 350 - 1500 350 - 1500 300 - 1500 200 - 1300 400 - 2000 500 - 1500 350 - 1900 300-2000 300-2400
mutant Tv-UV-5.6 and Tv-UV-5.9 primed two new bands of the same size (1100 and 1200 bp) and faintly primed in chemical mutants but absent in wild strain. Because of these genetic changes, these mutants might have
19594
Afr. J. Biotechnol.
A
M
1
B
2
3
4
5
1
2
3
4
C
5
M
Band 1 Band 4
Figure 2. RAPD DNA fragments amplified with decamer primers M-03 (A), M-07; (B), showing monomorphism and polymorphism; (C) DNA marker. Lane M indicates DNA marker and lane 1 indicates T. viride FCBP-142; lane 2, Tv-UV5.6; lane 3, Tv-UV-5.9; lane 4, Tv-Ch-4.3 and lane 5, Tv-Ch-5.5.
Figure 3. -RAPD profiles of T. viride FCBP-142 genotypes.
exhibited different production profiles. Band 4 (1000 bp) from the top in wild strain was mildly expressed, while it was overexpressed in the mutants. Chemical mutants were quite different from UV mutants and wild strain, and because of the significant differences in expression profile of different bands by different primers used, the
mutants might have outperformed the wild organism in enzyme production. Pattern of DNA fragments among genotypes granted a base to divide them in to major groups of genetic distance. Genotype-wise detail of number of polymorphic DNA fragments detected is given in Figure 3. On the basis of polymorphic DNA fragments,
Shafique et al.
19595
Figure 4. Homology tree constructed by an un-weighted pair group with arithmetic averages clustering algorithm from the pair wise matrix of genetic similarity amongst genotypes of T. viride FCBP-142.
genetic distance among T. viride genotypes was analyzed and it was estimated that the number of polymorphic DNA fragments ranged from 99 to 106. A homology tree was constructed by an un-weighted pair group method with arithmetic averages clustering algorithm from the pair wise matrix of genetic similarity among wild type and mutant genotypes (Figure 4). Different clusters of tree indicated a clear pattern of division among the genotypes. Two main groups of clusters were identified in the homology tree. First group comprised of three genotypes which included T. viride FCBP-142 (parental), Tv-UV-5.6 and Tv-UV-5.9. In these group genotypes Tv-UV-5.6 and Tv-UV-5.9 exhibited more genetic similarity as compared to other. The results therefore indicated that T. viride 142 had relatively acute genetic distance from the genotypes of its sub-cluster. In addition, the second major group consisted of the remaining two genotypes, Tv-Ch-4.3 and Tv-Ch-5.5, showing low genetic diversity among them. To further quantify the similarity between fungal isolates, the number of shared RAPD bands from all primers between each pair of isolates was compared. The five isolates from T. viride genotypes shared an average of 75% of bands (ranged= 54-95 %). Stability of the most efficient mutant derivatives of T. viride FCBP-142 (Tv-UV-5.6 and Tv-Ch-4.3) was evaluated up to 10 generations after every two months to check their ability of enzyme synthesis under adjusted assay conditions. It was
revealed from the results that these improved isolates exhibited same yield with insignificant difference up to tested generations (Figure 5). DISCUSSION Presently, the fungal strain T. viride FCBP-142 was mutagenized and genetically modified to develop a mutant strain capable of exhibiting high levels of cellulase activity because fungal strains have a unique character to pass over the environmental stress including chemical and irradiative mutagenesis and are highly susceptible to various physical as well as chemical mutagenic agents. This practice has become a routine in the field of biotechnology to develop a mutant through random mutagenesis (Azin and Noroozi, 2001; Mohsin, 2006). In the present study the induction of mutation through UV irradiation promoted the cellulase activity roughly about 2-fold. The improvement in enzyme activity may be due to the photolysis of pyramidines in adjacent pyramidines to form dimmers. They may cause error at the next replication and so result in mutation. The gene responsible for the production of cellulase may be over expressed due to mutation, as a result increase in enzyme activity (Gaedner et al., 1991; Chand et al., 2005). The main effect of mutagenic agents (x–rays, UV– rays, nitrous acid, dimethyl sulfonate, ethyl methane
Afr. J. Biotechnol.
(U/ml)
19596
Figure 5. Stability of Tv-UV-5.6 and Tv-Ch-4.3 mutant derivatives of T. viride FCBP-142 for cellulase production.
sulfonate (EMS) and acridine mustards) is to induce a lesion in or modification of the base sequence of DNA molecule; a mutation appears if this lesion remains un– repaired (Devehand and Gwynne, 1991). The perusal of data obtained from chemical treatment revealed that EMS proved more effective as it enhanced cellulase activity to the greater extent as compared to UV. This might be due to the fact that EMS is strong mutagenic agent and induces permanent changes in DNA structure (frame shift mutation). Similar research was conducted by Hamad et al. (2001), where they reported that chemical treatment is more efficient in inducing high-level mutations as compared to UV irradiation. Likewise, Mohsin (2006) worked on strain improvement of thermophilic fungi (Humicola insolens) by using both physical and chemical mutagens. The selection of efficient mutant strains was based on (qualitative) plate screening assay followed by (quantitative) submerged fermentation and compared with wild type strains. In this study, RAPD technique was used as a valuable diagnostic DNA marker system to evaluate genetic diversity of mutant genotypes and their wild types. It was revealed from the results that chemical mutants were significantly different from wild types and UV mutants with respect to their genetic make-up. Due to
different expression in these mutants, they might have given high yield of productivity during growth on different carbon sources. In several studies, RAPD fingerprinting technique has been employed to detect mutation, genetic relatedness and genetic variation within and between natural bacterial and human DNA and fungal populations (McDermott et al., 1994; Keinath et al., 1995). In a parallel study Larissa et al. (2002) applied RAPD fingerprinting to detect genetic variability among fourteen isolates of Trichoderma (six of T. viride, six of T. harzianum, one T. polysporum, one of T. pseudokoningii). Recently Chakraborty et al. (2010) studied the genetic relatedness among eleven isolates of T. viride and eight isolates of T. harzianum with six random primers. RAPD profiles showed genetic diversity among the isolates with the formation of eight clusters. Analysis of dendrogram revealed that similarity coefficient ranged from 0.67 to 0.95. Conclusion These study provided insight into improvement in enzyme production and the CMC hydrolysis by mutation of T. viride FCBP-142, thus saving the economy of the country and meeting industrial sector demand. The stable UV and
Shafique et al.
chemical mutants of T. viride obtained produced 2- and 3-fold more enzyme, respectively than produced by the wild-type organism. It is therefore recommended that RAPD is an efficient tool to detect genetic diversity at molecular level among genotypes, which were difficult to discriminate otherwise on the basis of morphological and physiological characters. REFERENCES Azin DA, Noroozi RF (2001). Effect of chemicals on the improved gluconate productivity by an Aspergillus niger strain. Appl. Biochem. Biotechnol. 61(3): 393-397. Beauchemin KA, Colombatto D, Morgavi DP, Yang WZ (2003). Use of exogenous fibrolytic enzymes to improve animal feed utilization by ruminants. J. Ani. Sci. 81(2): 37-47. Chakraborty BN, Chakraborty U, Saha A, Dey PL, Sunar K (2010). Molecular Characterization of Trichoderma viride and Trichoderma harzianum Isolated from Soils of North Bengal Based on rDNA Markers and Analysis of Their PCR-RAPD Profiles. Glob. J. Biotechnol. Biochem. 5(1): 55-61. Chand P, Aruna A, Maqsood AM, Rao LV (2005). Novel mutation method for increased cellulase production. J. App. Microbiol. 98(2): 318-323. Devehand M, Gwynne DI (1991). Expression of heterologous proteins in Aspergillus. J. Biotechnol. 17: 3-10. th Gaedner JE, Simmons JE, Snustad DP (1991). Principle of Genetics. 8 Ed. John Wiley and Sons, Inc. 304-305. Hamad A, Haq I, Qadeer MA, Javed I (2001). Screening of Bacillus licheniformis mutants for improved production of alpha-amylase. Pak. Jour. Bot. 33(special issue): 517-525. Jones C, Kortenkamp A (2000). RAPD library fingerprinting of bacterial and human DNA: Applications in mutation detection. Teratogen. Carcinogen. Mutagene. 20(2): 49-63. Keinath AP, Farnham MW, Zitter TA (1995). Morphological, pathological, genetic differentiation of Didymella bryoniae and Phoma species isolated from cucurbits. Phytopathol. 85: 364-369. Kotchoni OS, Shonukan OO (2002). Regulatory mutations affecting the synthesis of cellulose in B. pumilus. World J. Microbiol. Biotechnol. 18: 487-491.
19597
McDermott JM, Brandle U, Dutty F, Haemmerli UA, Keller S, Muller KE, Wolfe MS (1994). Genetic variation in powdery mildew of barly: Development of RAPD, Scar and VNTR markers. Phytopathol. 84: 1316-1321. Miller GL (1959). Use of dinitrosalicylic acid reagent for determination of reducing sugar. Ana. Chem. 31: 426-428. Minitab (2004). Multivariate cluster analysis. Compact disc software version. 13.0, USA. Mohsin MJ (2006). Improvement of a thermophilic fungal strain for cellulase production by chemical and UV mutagenesis, Ph. D. thesis, Dept. of Bot. GC Univ. Lahore. Ph. D. thesis. Morikawa Y, Kawamori M, Ado I, Shinsha I, Oda F, Takasawasa S, (1985). Improvement of cellulase production in Trichoderma reesei. Agric. Bio. Chem. 49: 1869. Nei N, Li W (1979). Mathematical model for studying genetic variation in terms of restriction endonucleases. Proc. Nat. Aca. Sci. USA 76: 5269-5273. Saghai-Maroof MA, Soliman KM, Jorgensen RA, Allard RW (1984). Fungal DNA Isolation. PNAS. 81: 8014-8018. Schulein M (2002). Protein engineering of cellulases: a review. Biochem. Biophy. Acta 1543: 239-252. Shafique S, Bajwa R, Shafique S (2009). Cellulase biosynthesis by selected Trichoderma species. Pak. J. Bot. 41(2): 907-916. Steel RGD, Torrie JH (1980). Principles and procedures of statistics, McGraw Hill Book Co., Inc. New York USA. Sun Y, Cheng J (2002). Hydrolysis of lignocellulosic material from ethanol production: a review. Biores. Tech. 83: 1-11. Williams JGK, Kubelik AR, Livak KJ, Rafalski JA, Tingey SV (1990). DNA polymorphisms amplified by arbitrary primers are useful as genetic markers. Nuc. Acids Res. 18: 6531-6535.
