Plant Molecular Biology 7:155-170 (1986) © Martinus N i j h o f f Publishers, Dordrecht - Printed in the Netherlands
155
Abscisic acid induction of cloned cotton late embryogenesis-abundant (Lea) m R N A s Glenn A. Galau, 1 D. Wayne Hughes I & Leon Dure III 2
1Department of Botany and 2Department of Biochemistry, University of Georgia, Athens, GA 30602, U.S.A. Keywords: embryo maturation, cDNA cloning, Gossypium hirsutum, hybrid-arrest translation, abscisic acid, embryo culture
Summary Earlier studies found that cotton (Gossypium hirsutum L.) cotyledons contain several mRNAs which are more abundant during late embryogenesis than in mid-embryogenesis or early germination. They are here termed 'Late embryogenesis-abundant' mRNAs, encoded by Lea loci. Complementary DNA clones for 18 such mRNA sequences, defined at a hybridization criterion of T m - 15 °C, were identified in a mature embryo cDNA library by differential cDNA hybridization. At a lower hybridization criterion, some sequence homology was found within several of these cloned Lea mRNA sequences. Each Lea mRNA sequence comprises 0.04-1.3% of mature embryo poly(A) ÷ mRNA, a level ten-fold to several hundred-fold higher than in young embryo or 24 h seedling poly(A)+mRNA. Of 18 Lea mRNA sequences examined in cultured young embryos, the level of at least 13 are specifically increased by exogenous abscisic acid (ABA), several to a level near that in normal mature embryos. However, the abundance of several o f the sequences does not appear to be significantly modulated by ABA. The LEA polypeptides encoded by 10 Lea mRNA sequences were identified by hybrid-arrested translation. They include most of the late embryogenesis-abundant~ ABAinducible, polypeptides previously identified. Preliminary results suggest that many o f the individual Lea mRNA sequences are transcribed from 1 - 3 genes in each of cotton's two subgenomes.
Introduction Cotton cotyledons contain several major mRNA species which are much more abundant in late maturation stage embryos than in younger embryos or young seedlings. These were identified both by cDNA-mRNA hybridization studies (17) and two dimensional gel analysis of polypeptides synthesized in vivo and in in vitro translation reactions (12). Of the above mRNAs whose translation products are detectable on two dimensional gels, all are found to be precociously induced and actively translated in excised young embryos if the embryos are incubated with abscisic acid at concentrations which inhibit precocious germination (12). Curi-
ously, no additional mRNAs were found to be similarly affected by ABA, indicating a very high correlation between an mRNA's inducibility by ABA and its maximum abundance occurring in mature embryos. Consequently it was proposed that endogenous ABA might regulate the concentration of these particular mRNAs during normal embryogenesis (12). Embryos of other species also contain putative ABA-regulated polypeptides and mRNAs, the regulation of which share some features with those in cotton (27). The only ABA-regulated mRNAs which have been studied with cloned probes are those for several storage proteins. These include the legumin (cruciferin) and napin of Brassica napis
Address correspondence to: Glenn A. Galau, Department of Botany, Miller Plant Sciences Building, University of Georgia, Athens,
GA 30602, U.S.A. Phone: (404) 542-3732. Telex: 810-754-3908/UGA LIBS ATHE.
156 L. (8), one of the vicilins (t3-conglycinin) of Glycine max (L.) Merr. (4), and the albumin E m storage protein of Triticum aestivum L. (33). Although these have been isolated and studied primarily for other reasons, all do share with the cotton m R N A s their specific inducibility by ABA in excised embryos. However, these particular storage protein m R N A s in soybean (4) and rapeseed (8), just as in cotton (13), decline precipitously in concentration prior to desiccation, and thus are not strictly analogous to those cotton or wheat m R N A s which are most abundant during this late stage of embryogenesis. Furthermore, unlike in wheat, soybean, or rapeseed embryos of comparable developmental stage, none of the three cotton storage protein m R N A s are regulated by ABA in excised embryos (12; Hughes and Galau, unpublished data). In further examination of the r o l e o f ABA in cotton embryogenesis, we have focused our attention on those cotton m R N A s which are most abundant in late embryogenesis. The cotton m R N A s described above have been earlier termed 'Subset 5' or 'ABA-inducible' (11, 12). Regardless of their possible regulation in other contexts, they are relatively abundant during late embryogenesis, and hence we use here the conservative term 'Late embryogenesis-abundant' (transcribed from Lea loci) to denote these m R N A sequences. These are operationally defined as those m R N A sequences which during normal embryogenesis and germination are significantly more abundant in maturation stage embryos than in younger embryos or young seedlings. This term does not imply their inducibility by ABA in normal or experimental embryogenesis/germination. To further address the issue of ABA inducibility of late embryogenesis-abundant m R N A sequences, as well as to initiate an examination of their structure and function, we report here the isolation and preliminary characterization of c D N A clones complementary to 18 different Lea m R N A sequences. Many encode the polypeptides identified earlier (12) and, of these, most are probably encoded by 1 - 3 genes in each of the two cotton subgenomes. At least 13 sequences are clearly inducible at the m R N A level by ABA in excised young embryos. However, at least one, and probably several, are probably not significantly regulated by ABA.
Materials and methods
Embryogenesis and embryo culture Cotton plants (Gossypium hirsutum L. cv Coker 201) were greenhouse-grown under natural light. Flowers were tagged the day of anthesis. Development of embryos takes 5 5 - 6 0 days until maturity, with cotyledons reaching a m a x i m u m fresh weight of 115+5 mg at 4 5 - 5 0 days. Young emrbyos were excised at 3 0 - 3 5 days post-anthesis, and had cotyledons with fresh weights of 50+5 mg. Only cotyledons were examined here. They are the same developmental stage as was examined in earlier studies (12, 13, 17, 18). Mature embryos were excised from dry seeds which had been stored at room temperature for up to one year. Such preparations contained traces of endosperm but, unless indicated otherwise, did not contain radicles. Normal germination of excised mature embryos proceeded at 28 °C in the dark between water-moistened filter paper for 24 h. Only cotyledons were used, and these prepartions are equivalent to the 24 h germination stage examined in the above earlier studies. Embryo culture will be described more fully elsewhere (Hughes, Bijaisoradat, and Galau, in preparation). Briefly, ovules containing young embryos were surface sterilized by brief immersion in 70070 ethanol and blotting on sterile paper towels. Using sterile gloved hands and forceps, embryos were gently removed from the ovules and individually washed two times each for 45 min in water. Hydrated endosperm was gently removed, and individual embryos placed in 20 ml glass scintillation vials containing 8 ml P G medium and 2070 sucrose, solifified with 0.4070 agarose (Biorad, regular low mr). P G medium contains 10 m M NHaNO 3, 5 m M KNO3, 2 m M CaCI 2, 0.7 m M KH2PO 4, 1.3 m M MgSO4, 25 #M Fe-EDTA, 5/zM Zn(NO3)2, 50 #M H3BO3, 50/zM MnSO4, 5/xM KI, 1 #M Na2MoO 4, 5 0 n M CuSO4, 5 0 n M CoC1 z, 4 # M nicotinic acid, 15 tzM thiamine • HCI, and 2.5/~M pyridoxine • HCI. All components, including sucrose, agarose, and ABA (when present), were autoclaved together at 1 × concentration after adjusting the p H to 5.5 with N a O H . Where embryos were to be exposed to ABA, mixed isomers of ABA (Sigma) were included in both the water washes and P G medium. Culture was at 2 9 - 3 0 °C in darkness. Vials were capped with 20 m m test tube closures and
157 sealed with parafilm. Under these conditions embryos will precociously germinate with radicle elongation and geotropism commencing within 12 h and unfolding of the cotyledons and hypocotyl elongation commencing within 24 h. Reversible inhibition of the initiation of germintion in 50% of embryos occurs in the presence o f 5 x 10 -7 M ABA and in 100% o f the embryos exposed to 1 0 - 6 - - 1 0 -4 M ABA.
R N A and D N A &olation Total RNA was isolated as described (19), and poly(A)+mRNA was isolated by chromatography on oligo(dT)-cellulose (2). Plasmid DNA was isolated, purified, and labeled with 32p by nicktranslation, all as described earlier (18). Radioactive insert DNA probes were prepared directly from crude plasmid DNA preparations (16) or from inserts isolated from restriction enzyme-digested DNA by sucrose gradient centrifugation or gel electrophoresis.
cDNA synthesis, cloning, and colonyfilterpreparation Synthesis of 32p-labeled cDNA complementary to poly(A)+mRNA, cloning of double-stranded cDNA into the Pst I site of pBR325 via dG:dC tails, transformation into SK1590, and preparation of bacterial colony-containing nitrocellulose filters were all as earlier described (18). Both the mature embryo cDNA library and the mature embryo cDNA used to screen this library was prepared using poly(A)+mRNA from radicle-containing embryos. About 3% o f the transformants contained nonrecombinant plasmids, based on their antibiotic resistances. Also, as indicated by colony hybridization with a lambda phage recombinant containing cotton rRNA genes (isolated by J. Kamalay), 2°70 o f the transformants contained cDNA inserts homologous to rRNA.
Electrophoresis of nucleic acids DNAs.were electrophoresed in agarose or polyacrylamide gels in Tris-borate/EDTA buffer (1), stained with ethidium bromide, photographed, and, where indicated, blotted onto nitrocellulose by the method of Southern (28). Inspection of the
filters with short wave UV light was used to confirm transfer and binding. For electrophoresis under denaturing conditions, 1/zg RNA was denatured in 5/zl 20 mM MOPSN a O H / 1 mM EDTA (pH 7.2), 50% deionized formamide, 6°70 formaldehyde, 0.02O7o bromphenol blue, and 5°7o sucrose at 60°C for l0 min and electrophoresed in submarine 1.7°70 agarose gels. Both the gel and electrophoresis buffer contained 40 mM M O P S - N a O H / 2 mM EDTA (pH 7.2) and 4O7o formaldehyde. The electrophoresis buffer was circulated between electrode chambers. For subsequent filter hybridization, the gel's top surface was scrubbed with 10x SSC ( l x SSC is 0.15 M NaCl, 0.015 M trisodium citrate) and blotted directly with 10x SSC, by the method of Southern (28), onto nylon membranes (Gene Screen, New England Nuclear). Before use the membranes were first boiled in 0.1x SPEP [1 x SPEP is 20 mM sodium phosphate buffer (pH 6.8), 10 mM EDTA-NaOH (pH 8.0), 1°70 SDS, 0.2o7o sodium pyrophosphate] and rinsed extensively in water and finally soaked in 10x SSC. After transfer, the filters were air dryed and the RNA cross-linked to the nylon with 500/~W • min/cm 2 (at 254 nm) illumination with a germicidal UV lamp (7). The RNA was visualized on the filters with short wave UV light, and the positions of rRNA marked on the filters as internal standards. Where several RNAs were to be compared, such inspection also confirmed that equal quantities of each RNA were present on the filters.
Filter hybridization Hybridization o f filters in aqueous buffers was essentially as described (18) but the number of post-hybridization washes was reduced. In experiments using nylon membranes the concentration of several o f the buffer components was increased several fold by their substitution by l x SPEP. Washing was always performed in the same buffer salts and at the same temperature as used during the hybridization. The Na ÷ concentration is specified in the text as SSC equivalents.
Hybrid-arrest translation To identify polypeptides encoded by cDNA insert-complementary mRNAs, a modification of the hybrid-arrest translation protocol (18, 26) was
158 used. Fifty/zg total RNA from radicle-containing mature embryos and 5/~g Pst I-cleaved plasmid D N A were dissolved in 50/zl 0.075 M PIPESN a O H (pH 6.8), 0.01 M EDTA-NaOH (pH 8.0), 0.35 M NaC1, 0.1070 SDS, 8007o deionized formamide. The nucleic acids were then denatured for 30 sec at 100°C, and incubated at 44°C or 50°C for 1 hr. Hybridization was quenched by the addition of cold ethanol and the nucleic acids collected and reprecipitated several times with ethanol. The nucleic acids were finally dissolved in 35 #I 5 m M H E P E S - K O H (pH 6.8). The sample was divided into two equal aliquots; one aliquot was treated at 100 °C for 1 min to denature any hybrids, and both aliquots were immediately quick frozen. The samples were later thawed by the addition of the wheat germ in vitro translation mixture. In vitro translation and the one and two dimensional electrophoresis of the [3H]leucine-labeled polypeptides were as described (10).
Evaluation of hybrization criteria Cotton nuclear D N A is approximately 3807o G + C (20, 30), and for simplicity cotton m R N A s are assumed to have approximately the same base composition. A compilation of diverse measurements in the literature (3, 6, 23, 29, 31) was used to estimate the melting temperatures of D N A - D N A and D N A - R N A hybrids in aqueous and formamide media. Corrections for formamide concentrations were those measured for D N A - R N A hybrids (6) and corrections for both Na ÷ concentration (31) and fragment length (5) were based on those of D N A - D N A duplexes in aqueous media. In summary, both cotton D N A - D N A duplexes and D N A - R N A hybrids of about 0 . 5 - 1 kb length are assumed to melt at 83°C, 91 °C, and 93°C in l x , 4 x , and 6 x SSC-equivalent aqueous buffers, respectively. Similar R N A - D N A hybrids are estimated to melt in 8007o formamide, 2× SSCequivalent (hybrid-arrest translation) buffer at 58°C. Results
Isolation of cDNA clones complementary to late embryogenesis-abundant m R N A s From earlier studies (12, 17), about
15 Lea
m R N A sequences were thought to each comprise, on the average, about 0.4°7o of the total mature embryo cotyledon poly(A)+mRNA, accounting together for about 6070 of the total m R N A mass. The concentration of these m R N A s were observed to be reduced at least 15-fold in cotyledons of both young embryos and in 24 h-germinated seedlings. The strategy for the isolation of c D N A clones complementary to these m R N A s was to screen a c D N A library, made from mature embryo m R N A , by hybridization with cDNAs made from the m R N A s of these three developmental stages. The relative intensity of hybridization of probe cDNAs to each colony should reflect the abundance of the cDNA's complement in the m R N A population used as probe (14, 25, 32), whereas the frequency of recovering a particular cloned sequence should reflect its abundance in the m R N A population cloned as cDNA. Double-stranded c D N A was synthesized from mature embryo poly(A)+mRNA, inserted into the Pst I site of pBR325 via dG:dC tailing, and transformed into E. coli. About 3 100 transformants were plated on filters in replicate and each replicate was hybridized in identical fashion with radioactive single-stranded c D N A made from cotyledon p o l y ( A ) + m R N A of young embryos, mature embryos, or 24 h-germinated seedlings (Fig. 1). Hybridization was at a moderate criterion in 4 x SSC-equivalent buffer at 68 °C, about 23 °C below the melting temperature ( T m - 2 3 ° C ) of wellmatched hybrids. Probes homologous to cotton r R N A and a storage protein c D N A insert from p G H S P C-72 (18) were also hybridized to colony replicates. O f the approximately 2500 colonies harboring recombinant plasmids containing m R N A sequences which were scorable for all three probe cDNAs, about 150, or 6070 of the total, had m R N A complements detectably more abundant in mature embryo p o l y ( A ) + m R N A (Fig. 1B) than in the p o l y ( A ) + m R N A of the other two stages (Fig. 1A and C). Most of these clones are a m o n g those circled in all panels of Fig. 1. Both the frequency of such c D N A clones and the developmental modulation of the concentrations of their complementary m R N A s were as expected if these were complementary to the Lea m R N A s described earlier. Ninety two putative Lea m R N A clones were selected for further study here. These were grown in liquid cul-
159
Fig. 1. Identification of colonies harboring Lea c D N A inserts. Replica filters containing 3 100 transformant colonies resulting from transformation with mature embryo cDNA insert-containing plasmids were individually hybridized with radioactive c D N A synthesized from poly(A) ÷ m R N A of young embryos (A), mature embryos (B), and 24 h-germinated seedlings (C). Hybridization was at a moderate criterion (T m - 2 3 °C) at 68 °C for 60 h in 4 x SSC-equivalent buffer containing 2 x 10s cpm 32p-labeled c D N A (3 x 108 cpm//zg) in 75 ml (0.1 ml/cm2). The autoradiograms shown here are 2 h exposures at - 8 0 °C with an intensifying screen. The circled colonies were later found to hybridize with a representative c D N A insert of one of the subsequently identified 18 Lea m R N A sequence groups, indicated by the number associated with each colony. Insert preparations of g2, g5, g6, g7, and g9 used in this study were from plasmids D132, D73, D29, both D7 and D49, and D152, respectively. The remaining inserts are specified in Table 1. Colonies identified in the Figure as C-72 and C-94 were added as internal standards and harbor plasmids p G H S P C-72 and p G H S P C-94, respectively. These contain about 1.8 kb c D N A inserts complementary to two cotton storage protein m R N A s . Open arrows indicate mature embryo cDNAcontaining colonies which hybridize with an insert preparation of p G H S P C-72.
ture, replica-plated onto filters, and rehybridized with the three developmental stage cDNAs, as before, with comparable results (data not shown).
Identification of individual Lea mRNA sequence groups Gossypium hirsutum is an allotetraploid (2n= 4 × = 52) with divergent A and D subgenomes as estimated by chromosome morphology and meiotic behavior (15). However, there is about 94°70 nucleotide sequence homology in at least 75°70 of the single-copy DNA of the two subgenomes (20, 30), the DNA fraction encoding the majority of the cotton cotyledon poly(A) + mRNAs (17). Homeologous gene transcripts are thus expected to be highly conserved. The extent of inbreeding of the cultivar is unknown, so allelic sequences might also be present in the population of selected Lea clones.
The strategy was to first identify those cDNA clones containing the same or very similar sequences, and to later explore the number of distinguishable transcripts within each sequence group and any possible homologies between sequence groups. A sequence group is defined as that group o f cDNA clones which cross-hybridize at a stringent ( T m - 1 5 ° C ) hybridization criterion in 1x SSC-equivalent buffer at 68°C. It was expected that allelic and homeologous gene transcripts would cross-hybridize at this criterion. In the following discussion, a particular sequence group is identified by the prefix 'g' and an arbitrary identification number. For simplicity, this nomenclature is used initially to identify the group of cDNA clones sharing the same sequence (operationally defined by cross-hybridization at T m - 1 5 ° C ) , particular mRNAs or cDNA clones in that sequence group, the polypeptides these encode, and the genes from which these mRNAs are transcribed.
160 Prior to cross-hybridization studies. The c D N A inserts contained in the selected plasmids were sized an agarose gels to ensure that all inserts were o f length sufficient to estimate their sequence homologies using this method. All the selected 92 plasmids had inserts in excess of 300 NT in length, with a number average length and (Gaussian) standard deviation of 720+180 NT. Initially, radioactive insert preparations made from crude D N A (16) of a total of 45 c D N A clones were individually hybridized at T m - 1 5 °C to replica filters containing the 92 selected colonies harboring Lea cDNAcontaining plasmids (data not shown). A total of 18 Lea sequence groups were identified by this method. Hybridization with any particular insert was to all or nearly all members of its sequence group. Due to occasional technical problems in filter or probe preparation, and possible insufficient sequence overlaps in some of the short c D N A insertcontaining clones, hybridization with more than one c D N A insert was required to unambiguously define the composition of some of the sequence groups. Eight sequence groups contained a single c D N A clone in the 92 member pool. The remaining groups contained 2 to 10 c D N A clones each, and a single group contained 29 c D N A clones (summarized in Table 1). From each sequence group containing more than one member in the selected pool, one or two plasmids containing the longest Pst I-liberatable insert were choosen as representatives for additional study. Conventionally prepared inserts of the representative clones of all sequence groups hybridized in stringent conditions with all the colonies in the 92 member-containing filters which were previously identified as belonging to the particular sequence group (data not shown). These same colonies (as well as others described below) on the original 3 100 colony-containing filters also hybridized with the individual sequence group representatives at high criterion (data not shown). Sequence homologies between Lea m R N A quence groups
se-
To investigate possible sequence homologies between sequence groups, insert preparations from representative plasmids of each sequence group were hybridized at lower hybridiztion criteria to colony blots containing all 92 selected bacterial
clones. Hybridization was at a permissive criterion of T m - 3 8 ° C in 6 x SSC-equivalent buffer at 55 °C or at an intermediate criterion of T m - 2 3 °C in 4 x SSC-equivalent buffer at 68°C. At the permissive criterion, several sequence groups showed unambiguous hybridization to bacterial colonies in addition to those known to contain the same sequence defined above at the stringent criterion of T m - 1 5 ° C (data not shown). Hybridization occured between colonies and representative probes of the following sequence groups: g2 (D132 as probe) and g4 (D19 as probe); g12 and g17; g6 (D29 as probe) and g7 (D7 as probe); and between g7 (D7 as probe) and gl0. Hybridization was reciprocol and involved most or all of the members of both sequence groups. When used as probe, g7 (D7 and D49 as probes) hybridized with at least 6 of 29 gl colonies (but not with the gl representative Dl13), and gl0 hybrididized with at least 4 of the 8 g6 colonies (but poorly with the g6 representatives D29 and D15). No hybridization to g7 or gl0 colonies was noted with insert probes for gl (Dl13 as probe) or g6 (D29 or D15 as probes), respectively. At the intermediate criterion, no cross hybridization was observed between any groups except g2 and g4. The colony blot results were confirmed and extended by more sensitive hybridization of the same insert preparations, at the permissive criterion, to filters containing electrophoretically-separated inserts of many of the cloned members of these crosshybridizing sequence groups (Fig. 2). This experiment shows that as might be expected, gl does have some homology with g6 and gl0, as well as with g7. The cross-hybridization patterns with different sequence group probes, however, suggest that not all cloned members of each group share equally the c o m m o n domain(s); that is, there may be more sequence differences evident within gl, g6, g7, and gl0 than might be expected on the basis of length differences alone. A possible caveat is cross reaction between the dG:dC tails contained on the c D N A inserts. Such hybridization probably is responsible for very minor, nonsequence groupspecific, cross reaction between about one-half of the selected 92 c D N A clones (data not shown), but is unlikely to be responsible for the homologies demonstrated here. Within the window of T m - 2 3 oC and T m - 3 8 °C, intergroup hybridization intensities ap-
161
"K
I
155
Abscisic acid induction of cloned cotton late embryogenesis-abundant (Lea) m R N A s Glenn A. Galau, 1 D. Wayne Hughes I & Leon Dure III 2
1Department of Botany and 2Department of Biochemistry, University of Georgia, Athens, GA 30602, U.S.A. Keywords: embryo maturation, cDNA cloning, Gossypium hirsutum, hybrid-arrest translation, abscisic acid, embryo culture
Summary Earlier studies found that cotton (Gossypium hirsutum L.) cotyledons contain several mRNAs which are more abundant during late embryogenesis than in mid-embryogenesis or early germination. They are here termed 'Late embryogenesis-abundant' mRNAs, encoded by Lea loci. Complementary DNA clones for 18 such mRNA sequences, defined at a hybridization criterion of T m - 15 °C, were identified in a mature embryo cDNA library by differential cDNA hybridization. At a lower hybridization criterion, some sequence homology was found within several of these cloned Lea mRNA sequences. Each Lea mRNA sequence comprises 0.04-1.3% of mature embryo poly(A) ÷ mRNA, a level ten-fold to several hundred-fold higher than in young embryo or 24 h seedling poly(A)+mRNA. Of 18 Lea mRNA sequences examined in cultured young embryos, the level of at least 13 are specifically increased by exogenous abscisic acid (ABA), several to a level near that in normal mature embryos. However, the abundance of several o f the sequences does not appear to be significantly modulated by ABA. The LEA polypeptides encoded by 10 Lea mRNA sequences were identified by hybrid-arrested translation. They include most of the late embryogenesis-abundant~ ABAinducible, polypeptides previously identified. Preliminary results suggest that many o f the individual Lea mRNA sequences are transcribed from 1 - 3 genes in each of cotton's two subgenomes.
Introduction Cotton cotyledons contain several major mRNA species which are much more abundant in late maturation stage embryos than in younger embryos or young seedlings. These were identified both by cDNA-mRNA hybridization studies (17) and two dimensional gel analysis of polypeptides synthesized in vivo and in in vitro translation reactions (12). Of the above mRNAs whose translation products are detectable on two dimensional gels, all are found to be precociously induced and actively translated in excised young embryos if the embryos are incubated with abscisic acid at concentrations which inhibit precocious germination (12). Curi-
ously, no additional mRNAs were found to be similarly affected by ABA, indicating a very high correlation between an mRNA's inducibility by ABA and its maximum abundance occurring in mature embryos. Consequently it was proposed that endogenous ABA might regulate the concentration of these particular mRNAs during normal embryogenesis (12). Embryos of other species also contain putative ABA-regulated polypeptides and mRNAs, the regulation of which share some features with those in cotton (27). The only ABA-regulated mRNAs which have been studied with cloned probes are those for several storage proteins. These include the legumin (cruciferin) and napin of Brassica napis
Address correspondence to: Glenn A. Galau, Department of Botany, Miller Plant Sciences Building, University of Georgia, Athens,
GA 30602, U.S.A. Phone: (404) 542-3732. Telex: 810-754-3908/UGA LIBS ATHE.
156 L. (8), one of the vicilins (t3-conglycinin) of Glycine max (L.) Merr. (4), and the albumin E m storage protein of Triticum aestivum L. (33). Although these have been isolated and studied primarily for other reasons, all do share with the cotton m R N A s their specific inducibility by ABA in excised embryos. However, these particular storage protein m R N A s in soybean (4) and rapeseed (8), just as in cotton (13), decline precipitously in concentration prior to desiccation, and thus are not strictly analogous to those cotton or wheat m R N A s which are most abundant during this late stage of embryogenesis. Furthermore, unlike in wheat, soybean, or rapeseed embryos of comparable developmental stage, none of the three cotton storage protein m R N A s are regulated by ABA in excised embryos (12; Hughes and Galau, unpublished data). In further examination of the r o l e o f ABA in cotton embryogenesis, we have focused our attention on those cotton m R N A s which are most abundant in late embryogenesis. The cotton m R N A s described above have been earlier termed 'Subset 5' or 'ABA-inducible' (11, 12). Regardless of their possible regulation in other contexts, they are relatively abundant during late embryogenesis, and hence we use here the conservative term 'Late embryogenesis-abundant' (transcribed from Lea loci) to denote these m R N A sequences. These are operationally defined as those m R N A sequences which during normal embryogenesis and germination are significantly more abundant in maturation stage embryos than in younger embryos or young seedlings. This term does not imply their inducibility by ABA in normal or experimental embryogenesis/germination. To further address the issue of ABA inducibility of late embryogenesis-abundant m R N A sequences, as well as to initiate an examination of their structure and function, we report here the isolation and preliminary characterization of c D N A clones complementary to 18 different Lea m R N A sequences. Many encode the polypeptides identified earlier (12) and, of these, most are probably encoded by 1 - 3 genes in each of the two cotton subgenomes. At least 13 sequences are clearly inducible at the m R N A level by ABA in excised young embryos. However, at least one, and probably several, are probably not significantly regulated by ABA.
Materials and methods
Embryogenesis and embryo culture Cotton plants (Gossypium hirsutum L. cv Coker 201) were greenhouse-grown under natural light. Flowers were tagged the day of anthesis. Development of embryos takes 5 5 - 6 0 days until maturity, with cotyledons reaching a m a x i m u m fresh weight of 115+5 mg at 4 5 - 5 0 days. Young emrbyos were excised at 3 0 - 3 5 days post-anthesis, and had cotyledons with fresh weights of 50+5 mg. Only cotyledons were examined here. They are the same developmental stage as was examined in earlier studies (12, 13, 17, 18). Mature embryos were excised from dry seeds which had been stored at room temperature for up to one year. Such preparations contained traces of endosperm but, unless indicated otherwise, did not contain radicles. Normal germination of excised mature embryos proceeded at 28 °C in the dark between water-moistened filter paper for 24 h. Only cotyledons were used, and these prepartions are equivalent to the 24 h germination stage examined in the above earlier studies. Embryo culture will be described more fully elsewhere (Hughes, Bijaisoradat, and Galau, in preparation). Briefly, ovules containing young embryos were surface sterilized by brief immersion in 70070 ethanol and blotting on sterile paper towels. Using sterile gloved hands and forceps, embryos were gently removed from the ovules and individually washed two times each for 45 min in water. Hydrated endosperm was gently removed, and individual embryos placed in 20 ml glass scintillation vials containing 8 ml P G medium and 2070 sucrose, solifified with 0.4070 agarose (Biorad, regular low mr). P G medium contains 10 m M NHaNO 3, 5 m M KNO3, 2 m M CaCI 2, 0.7 m M KH2PO 4, 1.3 m M MgSO4, 25 #M Fe-EDTA, 5/zM Zn(NO3)2, 50 #M H3BO3, 50/zM MnSO4, 5/xM KI, 1 #M Na2MoO 4, 5 0 n M CuSO4, 5 0 n M CoC1 z, 4 # M nicotinic acid, 15 tzM thiamine • HCI, and 2.5/~M pyridoxine • HCI. All components, including sucrose, agarose, and ABA (when present), were autoclaved together at 1 × concentration after adjusting the p H to 5.5 with N a O H . Where embryos were to be exposed to ABA, mixed isomers of ABA (Sigma) were included in both the water washes and P G medium. Culture was at 2 9 - 3 0 °C in darkness. Vials were capped with 20 m m test tube closures and
157 sealed with parafilm. Under these conditions embryos will precociously germinate with radicle elongation and geotropism commencing within 12 h and unfolding of the cotyledons and hypocotyl elongation commencing within 24 h. Reversible inhibition of the initiation of germintion in 50% of embryos occurs in the presence o f 5 x 10 -7 M ABA and in 100% o f the embryos exposed to 1 0 - 6 - - 1 0 -4 M ABA.
R N A and D N A &olation Total RNA was isolated as described (19), and poly(A)+mRNA was isolated by chromatography on oligo(dT)-cellulose (2). Plasmid DNA was isolated, purified, and labeled with 32p by nicktranslation, all as described earlier (18). Radioactive insert DNA probes were prepared directly from crude plasmid DNA preparations (16) or from inserts isolated from restriction enzyme-digested DNA by sucrose gradient centrifugation or gel electrophoresis.
cDNA synthesis, cloning, and colonyfilterpreparation Synthesis of 32p-labeled cDNA complementary to poly(A)+mRNA, cloning of double-stranded cDNA into the Pst I site of pBR325 via dG:dC tails, transformation into SK1590, and preparation of bacterial colony-containing nitrocellulose filters were all as earlier described (18). Both the mature embryo cDNA library and the mature embryo cDNA used to screen this library was prepared using poly(A)+mRNA from radicle-containing embryos. About 3% o f the transformants contained nonrecombinant plasmids, based on their antibiotic resistances. Also, as indicated by colony hybridization with a lambda phage recombinant containing cotton rRNA genes (isolated by J. Kamalay), 2°70 o f the transformants contained cDNA inserts homologous to rRNA.
Electrophoresis of nucleic acids DNAs.were electrophoresed in agarose or polyacrylamide gels in Tris-borate/EDTA buffer (1), stained with ethidium bromide, photographed, and, where indicated, blotted onto nitrocellulose by the method of Southern (28). Inspection of the
filters with short wave UV light was used to confirm transfer and binding. For electrophoresis under denaturing conditions, 1/zg RNA was denatured in 5/zl 20 mM MOPSN a O H / 1 mM EDTA (pH 7.2), 50% deionized formamide, 6°70 formaldehyde, 0.02O7o bromphenol blue, and 5°7o sucrose at 60°C for l0 min and electrophoresed in submarine 1.7°70 agarose gels. Both the gel and electrophoresis buffer contained 40 mM M O P S - N a O H / 2 mM EDTA (pH 7.2) and 4O7o formaldehyde. The electrophoresis buffer was circulated between electrode chambers. For subsequent filter hybridization, the gel's top surface was scrubbed with 10x SSC ( l x SSC is 0.15 M NaCl, 0.015 M trisodium citrate) and blotted directly with 10x SSC, by the method of Southern (28), onto nylon membranes (Gene Screen, New England Nuclear). Before use the membranes were first boiled in 0.1x SPEP [1 x SPEP is 20 mM sodium phosphate buffer (pH 6.8), 10 mM EDTA-NaOH (pH 8.0), 1°70 SDS, 0.2o7o sodium pyrophosphate] and rinsed extensively in water and finally soaked in 10x SSC. After transfer, the filters were air dryed and the RNA cross-linked to the nylon with 500/~W • min/cm 2 (at 254 nm) illumination with a germicidal UV lamp (7). The RNA was visualized on the filters with short wave UV light, and the positions of rRNA marked on the filters as internal standards. Where several RNAs were to be compared, such inspection also confirmed that equal quantities of each RNA were present on the filters.
Filter hybridization Hybridization o f filters in aqueous buffers was essentially as described (18) but the number of post-hybridization washes was reduced. In experiments using nylon membranes the concentration of several o f the buffer components was increased several fold by their substitution by l x SPEP. Washing was always performed in the same buffer salts and at the same temperature as used during the hybridization. The Na ÷ concentration is specified in the text as SSC equivalents.
Hybrid-arrest translation To identify polypeptides encoded by cDNA insert-complementary mRNAs, a modification of the hybrid-arrest translation protocol (18, 26) was
158 used. Fifty/zg total RNA from radicle-containing mature embryos and 5/~g Pst I-cleaved plasmid D N A were dissolved in 50/zl 0.075 M PIPESN a O H (pH 6.8), 0.01 M EDTA-NaOH (pH 8.0), 0.35 M NaC1, 0.1070 SDS, 8007o deionized formamide. The nucleic acids were then denatured for 30 sec at 100°C, and incubated at 44°C or 50°C for 1 hr. Hybridization was quenched by the addition of cold ethanol and the nucleic acids collected and reprecipitated several times with ethanol. The nucleic acids were finally dissolved in 35 #I 5 m M H E P E S - K O H (pH 6.8). The sample was divided into two equal aliquots; one aliquot was treated at 100 °C for 1 min to denature any hybrids, and both aliquots were immediately quick frozen. The samples were later thawed by the addition of the wheat germ in vitro translation mixture. In vitro translation and the one and two dimensional electrophoresis of the [3H]leucine-labeled polypeptides were as described (10).
Evaluation of hybrization criteria Cotton nuclear D N A is approximately 3807o G + C (20, 30), and for simplicity cotton m R N A s are assumed to have approximately the same base composition. A compilation of diverse measurements in the literature (3, 6, 23, 29, 31) was used to estimate the melting temperatures of D N A - D N A and D N A - R N A hybrids in aqueous and formamide media. Corrections for formamide concentrations were those measured for D N A - R N A hybrids (6) and corrections for both Na ÷ concentration (31) and fragment length (5) were based on those of D N A - D N A duplexes in aqueous media. In summary, both cotton D N A - D N A duplexes and D N A - R N A hybrids of about 0 . 5 - 1 kb length are assumed to melt at 83°C, 91 °C, and 93°C in l x , 4 x , and 6 x SSC-equivalent aqueous buffers, respectively. Similar R N A - D N A hybrids are estimated to melt in 8007o formamide, 2× SSCequivalent (hybrid-arrest translation) buffer at 58°C. Results
Isolation of cDNA clones complementary to late embryogenesis-abundant m R N A s From earlier studies (12, 17), about
15 Lea
m R N A sequences were thought to each comprise, on the average, about 0.4°7o of the total mature embryo cotyledon poly(A)+mRNA, accounting together for about 6070 of the total m R N A mass. The concentration of these m R N A s were observed to be reduced at least 15-fold in cotyledons of both young embryos and in 24 h-germinated seedlings. The strategy for the isolation of c D N A clones complementary to these m R N A s was to screen a c D N A library, made from mature embryo m R N A , by hybridization with cDNAs made from the m R N A s of these three developmental stages. The relative intensity of hybridization of probe cDNAs to each colony should reflect the abundance of the cDNA's complement in the m R N A population used as probe (14, 25, 32), whereas the frequency of recovering a particular cloned sequence should reflect its abundance in the m R N A population cloned as cDNA. Double-stranded c D N A was synthesized from mature embryo poly(A)+mRNA, inserted into the Pst I site of pBR325 via dG:dC tailing, and transformed into E. coli. About 3 100 transformants were plated on filters in replicate and each replicate was hybridized in identical fashion with radioactive single-stranded c D N A made from cotyledon p o l y ( A ) + m R N A of young embryos, mature embryos, or 24 h-germinated seedlings (Fig. 1). Hybridization was at a moderate criterion in 4 x SSC-equivalent buffer at 68 °C, about 23 °C below the melting temperature ( T m - 2 3 ° C ) of wellmatched hybrids. Probes homologous to cotton r R N A and a storage protein c D N A insert from p G H S P C-72 (18) were also hybridized to colony replicates. O f the approximately 2500 colonies harboring recombinant plasmids containing m R N A sequences which were scorable for all three probe cDNAs, about 150, or 6070 of the total, had m R N A complements detectably more abundant in mature embryo p o l y ( A ) + m R N A (Fig. 1B) than in the p o l y ( A ) + m R N A of the other two stages (Fig. 1A and C). Most of these clones are a m o n g those circled in all panels of Fig. 1. Both the frequency of such c D N A clones and the developmental modulation of the concentrations of their complementary m R N A s were as expected if these were complementary to the Lea m R N A s described earlier. Ninety two putative Lea m R N A clones were selected for further study here. These were grown in liquid cul-
159
Fig. 1. Identification of colonies harboring Lea c D N A inserts. Replica filters containing 3 100 transformant colonies resulting from transformation with mature embryo cDNA insert-containing plasmids were individually hybridized with radioactive c D N A synthesized from poly(A) ÷ m R N A of young embryos (A), mature embryos (B), and 24 h-germinated seedlings (C). Hybridization was at a moderate criterion (T m - 2 3 °C) at 68 °C for 60 h in 4 x SSC-equivalent buffer containing 2 x 10s cpm 32p-labeled c D N A (3 x 108 cpm//zg) in 75 ml (0.1 ml/cm2). The autoradiograms shown here are 2 h exposures at - 8 0 °C with an intensifying screen. The circled colonies were later found to hybridize with a representative c D N A insert of one of the subsequently identified 18 Lea m R N A sequence groups, indicated by the number associated with each colony. Insert preparations of g2, g5, g6, g7, and g9 used in this study were from plasmids D132, D73, D29, both D7 and D49, and D152, respectively. The remaining inserts are specified in Table 1. Colonies identified in the Figure as C-72 and C-94 were added as internal standards and harbor plasmids p G H S P C-72 and p G H S P C-94, respectively. These contain about 1.8 kb c D N A inserts complementary to two cotton storage protein m R N A s . Open arrows indicate mature embryo cDNAcontaining colonies which hybridize with an insert preparation of p G H S P C-72.
ture, replica-plated onto filters, and rehybridized with the three developmental stage cDNAs, as before, with comparable results (data not shown).
Identification of individual Lea mRNA sequence groups Gossypium hirsutum is an allotetraploid (2n= 4 × = 52) with divergent A and D subgenomes as estimated by chromosome morphology and meiotic behavior (15). However, there is about 94°70 nucleotide sequence homology in at least 75°70 of the single-copy DNA of the two subgenomes (20, 30), the DNA fraction encoding the majority of the cotton cotyledon poly(A) + mRNAs (17). Homeologous gene transcripts are thus expected to be highly conserved. The extent of inbreeding of the cultivar is unknown, so allelic sequences might also be present in the population of selected Lea clones.
The strategy was to first identify those cDNA clones containing the same or very similar sequences, and to later explore the number of distinguishable transcripts within each sequence group and any possible homologies between sequence groups. A sequence group is defined as that group o f cDNA clones which cross-hybridize at a stringent ( T m - 1 5 ° C ) hybridization criterion in 1x SSC-equivalent buffer at 68°C. It was expected that allelic and homeologous gene transcripts would cross-hybridize at this criterion. In the following discussion, a particular sequence group is identified by the prefix 'g' and an arbitrary identification number. For simplicity, this nomenclature is used initially to identify the group of cDNA clones sharing the same sequence (operationally defined by cross-hybridization at T m - 1 5 ° C ) , particular mRNAs or cDNA clones in that sequence group, the polypeptides these encode, and the genes from which these mRNAs are transcribed.
160 Prior to cross-hybridization studies. The c D N A inserts contained in the selected plasmids were sized an agarose gels to ensure that all inserts were o f length sufficient to estimate their sequence homologies using this method. All the selected 92 plasmids had inserts in excess of 300 NT in length, with a number average length and (Gaussian) standard deviation of 720+180 NT. Initially, radioactive insert preparations made from crude D N A (16) of a total of 45 c D N A clones were individually hybridized at T m - 1 5 °C to replica filters containing the 92 selected colonies harboring Lea cDNAcontaining plasmids (data not shown). A total of 18 Lea sequence groups were identified by this method. Hybridization with any particular insert was to all or nearly all members of its sequence group. Due to occasional technical problems in filter or probe preparation, and possible insufficient sequence overlaps in some of the short c D N A insertcontaining clones, hybridization with more than one c D N A insert was required to unambiguously define the composition of some of the sequence groups. Eight sequence groups contained a single c D N A clone in the 92 member pool. The remaining groups contained 2 to 10 c D N A clones each, and a single group contained 29 c D N A clones (summarized in Table 1). From each sequence group containing more than one member in the selected pool, one or two plasmids containing the longest Pst I-liberatable insert were choosen as representatives for additional study. Conventionally prepared inserts of the representative clones of all sequence groups hybridized in stringent conditions with all the colonies in the 92 member-containing filters which were previously identified as belonging to the particular sequence group (data not shown). These same colonies (as well as others described below) on the original 3 100 colony-containing filters also hybridized with the individual sequence group representatives at high criterion (data not shown). Sequence homologies between Lea m R N A quence groups
se-
To investigate possible sequence homologies between sequence groups, insert preparations from representative plasmids of each sequence group were hybridized at lower hybridiztion criteria to colony blots containing all 92 selected bacterial
clones. Hybridization was at a permissive criterion of T m - 3 8 ° C in 6 x SSC-equivalent buffer at 55 °C or at an intermediate criterion of T m - 2 3 °C in 4 x SSC-equivalent buffer at 68°C. At the permissive criterion, several sequence groups showed unambiguous hybridization to bacterial colonies in addition to those known to contain the same sequence defined above at the stringent criterion of T m - 1 5 ° C (data not shown). Hybridization occured between colonies and representative probes of the following sequence groups: g2 (D132 as probe) and g4 (D19 as probe); g12 and g17; g6 (D29 as probe) and g7 (D7 as probe); and between g7 (D7 as probe) and gl0. Hybridization was reciprocol and involved most or all of the members of both sequence groups. When used as probe, g7 (D7 and D49 as probes) hybridized with at least 6 of 29 gl colonies (but not with the gl representative Dl13), and gl0 hybrididized with at least 4 of the 8 g6 colonies (but poorly with the g6 representatives D29 and D15). No hybridization to g7 or gl0 colonies was noted with insert probes for gl (Dl13 as probe) or g6 (D29 or D15 as probes), respectively. At the intermediate criterion, no cross hybridization was observed between any groups except g2 and g4. The colony blot results were confirmed and extended by more sensitive hybridization of the same insert preparations, at the permissive criterion, to filters containing electrophoretically-separated inserts of many of the cloned members of these crosshybridizing sequence groups (Fig. 2). This experiment shows that as might be expected, gl does have some homology with g6 and gl0, as well as with g7. The cross-hybridization patterns with different sequence group probes, however, suggest that not all cloned members of each group share equally the c o m m o n domain(s); that is, there may be more sequence differences evident within gl, g6, g7, and gl0 than might be expected on the basis of length differences alone. A possible caveat is cross reaction between the dG:dC tails contained on the c D N A inserts. Such hybridization probably is responsible for very minor, nonsequence groupspecific, cross reaction between about one-half of the selected 92 c D N A clones (data not shown), but is unlikely to be responsible for the homologies demonstrated here. Within the window of T m - 2 3 oC and T m - 3 8 °C, intergroup hybridization intensities ap-
161
"K
I
