Plant Molecular Biology 29: 841-856, 1995. © 1995 Kluwer Academic Publishers. Printed in Belgium. 841 Expression of heat shock factor and heat shock protein 70 genes during maize pollen development Dominique Gagliardi 1, Christian Breton, Annie Chaboud, Philippe Vergne* and Christian Dumas Ecole Normale Supdrieure de Lyon, Reconnaissance Cellulaire et Amdlioration des Plantes, UMR CNRS-INRA 9938, 46 Allde d'Italie, 69364 Lyon cedex 07, France (* author for correspondence); 1present address: Department of Plant Sciences, University of Oxford, South Parks Road, Oxford, OX1 3RB, UK Received 27 February 1995; accepted in revised form 18 August 1995 Key words: development, heat shock factor, heat shock protein, maize, pollen Abstract We have analysed the expression of heat shock protein 70 (HSP70) and heat shock factor (HSF) gene during maize pollen development, HSFs being the transcriptional activators of hsp genes. In order to eliminate the sporophytic tissues of anthers, we have isolated homogeneous cell populations corresponding to five stages of maize pollen development from microspores to mature pollen. We show that in the absence of heat stress, hsp70 genes are highly expressed late-bicellular pollen as compared to other stages. HSP70 transcripts are significantly accumulated in response to a heat shock at the late microspore stage but to a much lower extent than in vegetative tissues. The latest stages of pollen development, i.e. mid-tricellular and mature pollen, do not exhibit heat-induced accumulation of HSP70 transcripts. Therefore, we analysed the expression of hsf genes throughout pollen development. We demonstrate that at least three hsf genes are expressed in maize and that transcripts corresponding to one hsf gene, whose expression is independent of temperature in somatic as well as in microgametophytic tis sues, are present at similar levels throughout pollen development. In addition, we show that the expression of the two other hsf genes is heat-inducible in maize vegetative tissues and is not significantly increased after heat shock at any stage of pollen development. These results indicate that the loss of hsp gene expression at late stages of pollen development is not due to a modification of hsf gene expression at the mRNA level and that hsf gene expression is differentially regulated in vegetative and microgametophytic tissues. Introduction Heat shock proteins (HSPs) constitute a nonhomogeneous group of proteins that have been originally defined by their increased expression after a temperature upshift [2, 51 ]. However, several members of the multigenic hsp gene families are also expressed during normal growth in the The nucleotide sequence data reported in this paper will appear in the EMBL, GenBank and DDBJ Nucleotide Sequence Databases under the accession number X82943. 842 absence of stress [2, 51]. During the past few years, transcription of members of hsp90, hsp70 and low-molecular-weight hsp gene families have been reported during gametogenesis and embryogenesis in animals [26]. However, the precise role of these HSPs in developmental decisions is still unknown. Interestingly, the heat-induced expression of inducible hsp genes is often repressed or incomplete during developmental processes [26]. In plants, the developmentally regulated expression of hsp genes has been mostly studied during embryogenesis, flower and anther development [15, 17, 18, 30, 38, 49, 51, 52]. In particular, it has been shown that members of hsp90 and hsp18 gene families are specifically expressed at certain developmental stages of whole maize anthers [4, 10, 19, 33]. However, it remains to be determined whether these genes are expressed within the sporophytic tissues of the anther or in the microgametophytes per se. The ability of the HSPs to be synthesized upon heat shock has been shown to be lost gradually during maize pollen development [23] and is completely lost in mature and germinating pollen grains [16, 22, 23]. Likewise, a reporter gene under the control of the hsp70 promoter of Drosophila is induced by heat in vegetative tissues of transgenic tobacco plants whereas heat shock fails to stimulate its expression in pollen [46]. However, in transcriptionally active germinating pollen, hsp70 and hsp18 genes are transcribed in response to heat shock, but at a very low level not sufficient for an effective accumulation of HSPs [27]. These studies on both developmentally regulated and heat-induced expression of hsp genes during pollen development raise the problem of the control of hsp gene transcription. In eukaryotes, transcriptional control of hsp genes is mediated by the heat shock factors (HSFs) [31, 36, 45]. HSF binds to the heat shock element (HSE), a conserved sequence present in multiple copies in the promoter regions of hsp genes. In animals and yeasts, a constitutively synthesized HSF is responsible for the activation of hsp genes upon heat shock, and hsf gene expression is not modified by the temperature increase. In contrast, in tomato, the transcription of two hsf genes is induced by heat shock while another one is constitutively expressed [42]. A hsf gene whose expression is increased upon heat-shock has also been recently characterized in Arabidopsis thaliana [28]. Hence, the expression of hsf genes seems to be regulated in a different manner in plants and animals. To date, the expression of hsf genes during differentiation or developmental processes has been investigated in mammals and chicken. Analysis of HSF activity in mammalian developmental systems have suggested that distinct HSFs are responsible for the stress-induced and the developmentally regulated expression of hsp genes [35, 36, 41]. No data are currently available concerning hsf gene expression during any plant development process. In the present paper, we have investigated the developmentally regulated and heat-induced expression of hsp70 and hsf genes during maize pollen development. To this end, we have isolated pollen at five stages of development in order to eliminate any contribution of the sporophytic tissues of the anthers. We show that hsp70 genes exhibit a developmentally regulated expression pattern and that HSP70 transcripts accumulate in response to heat shock only at the beginning of pollen development but to a much lesser extent than in vegetative tissues. Our results also indicate that transcripts of one hsf gene are present throughout pollen development, including mature pollen and that the expression of two other hsf genes, which is heat-inducible in vegetative tissues, is not significantly enhanced upon heat shock at any stage of pollen development. These data lead to the conclusion that although hsf genes appeared to be differentially regulated in vegetative and microgametophytic tissues, the loss of hsp gene activation during maize pollen development is not due to a modification of the expression of hsf genes. Material and methods Plant material Maize (Zea mays L.) plants (genotype DH5 × DH7) were grown with a 16 h illumination period 843 ( 7 0 0 - 8 0 0 / ~ E m - 2 s -1) at 2 4 / 1 9 + 1 ° C (day/ night) and 80 + 5~o relative humidity. DH5 and DH7 are doubled haploid fines obtained by in vitro androgenesis [ 5 ]. For vegetative tissue analysis, 10-day old plantlets were incubated for 2 h at 28 °C or 40 °C in water-saturated atmosphere. Leaves and roots were then quickly excised and immediately frozen in liquid nitrogen. Immature pollen isolation Five different stages of pollen development were studied: late microspore (vacuolated), late-bicellular pollen (approximately half starch-filled), early-tricellular pollen, mid-tricellular pollen and mature pollen. Late microspores were isolated as described by GaiUard et al. [24]. Late-bicellular and early-tricellular pollen were isolated from spikelets located at the bottom of fully emerged tassels which did not bear any dehiscing anthers. These selected anthers would have produced mature pollen 7-8 days later. Fragments of tassel branches were submerged in 0.5~o (w/v) sodium hypochlorite solution for 10min at 4 °C. The whole isolation procedure was then performed at 4 °C under sterile conditions. After 5 rinses with cold sterile water, each fragment was cut into 1 - 2 m m pieces in MBKS12 buffer (20mM MOPS pH 7.1, Brewbacker and Kwack's salts [12], 12~o sucrose (w/v)). After gentle stirring, the slurry was filtered through a 100/~m pore size sieve to remove large anther debris. Immature pollen was then retained on a 50/~M pre size sieve. After resuspension in MBKS 12, bicellular and early-tricellular pollen were loaded onto discontinuous Percoll (Pharmacia) gradients (25 ~o, 35 ~o and 50 ~ (v/v) in MBKS 12) and centrifuged at 250 × g for 4 min. Early-tricellular pollen pelleted to the bottom of the gradient. The bicellular pollen fraction was collected at the top of the 35 ~o Percoll layer and, in order to eliminate some contaminating stages, recentrifuged at 250 × g for 4 min through a 25-35~o Percoll gradient. Both late-bicellular and early-tricellular pollen fractions were collected and washed once with MBKS12 and incubated in this buffer as described below. Mid-tricellular pollen was isolated as detailed for early-tricellular pollen but from the bottom of tassels whose tops bore dehiscing anthers. The selected anthers would have produced pollen 3-4 days later. Mature pollen was collected at anthesis and it was checked that its water content, which is a good criterion for pollen quality [29], was at least 55 ~/o (w/w) of its fresh weight. Incubation buffer for mature pollen was 20 mM MOPS pH 7.1, Brewbaker and Kwack's salts [ 12], 15 ~ sucrose (w/v). All microgametophyte populations were incubated for 2 h at 28 °C or 40 °C and then stored in liquid nitrogen until analysis. Nucleic acid extraction Poly(A) + RNA was extracted using Dynabeads oligo(dT)25 (Dynal) according to manufacturer's recommendations. Genomia DNA was extracted using the method of Rogers and Bendich [40] from 10-day old plantlet leaves. Production of a maize hsf probe Unstressed leaf poly(A) ÷ RNA (0.5#g) was treated by DNaseI (Boehringer-Mannheim) and extracted once with phenol-chloroform and twice with chloroform. After ethanol precipitation, single-strand cDNA was synthesized for 1.5 h at 42 °C in a 50/~1 reaction of 50 mM Tris-HCl pH 8.3, 75 mM KCI, 3 mM MgC12, 40 U RNasin, 10mM DTT, 0.5raM dNTP, 20gg/ml oligo(dT)16_18, 0.1mg/ml BSA and 200U M-MLV reverse transcriptase (Gibco BRL). Sense (P1) and antisense (P2) PCR primers were derived from the amino acid sequences P(KR)(YF)FKH and W(EQ)F(AE)NE which correspond to conserved motifs in the DNA binding domain of tomato, Arabidopsis and yeast HSFs (see Fig. 1 and Fig. 4 for primer sequences and motif positions). In a reaction volume of 100 #1, 2~o of an unstressed leaf cDNA reaction was used for amplification in 10 mM Tris-HCl pH 8.3, 50 mM KC1, 2 mM MgC12, 200 #M of 844 each dNTP, 1 #M of each primer and 2.5 U of AmpliTaq DNA polymerase (Perkin-Elmer Cetus). PCR conditions were 94 °C/5 min followed by 30 cycles of 94 °C/45 s; 50 °C/45 s; 72 0C/45 s and afinal step at 72 °C/5 min. Of the reaction 1 ~o was subjected to a second round of PCR amplification under similar conditions with primers corresponding to P1 and P2 but containing Pst I and Cla I restriction sites, respectively. The resulting fragment was cloned into pBluescript and sequenced by the dideoxy chain termination method (Sequenase 2.0 kit, US Biochemical). This 101 bp fragment is further referred to as 'hsf probe' (see Results). Construction and screening of cDNA libraries Two maize (inbred line A188) PCR-generated cDNA libraries obtained from male gametes (sperm cells) and transition-stage embryos were screened using the 'hsf probe'. The construction of the maize transition-stage embryo cDNA library is described elsewhere [11]. Maize male gamete-enriched fractions were prepared from mature pollen according to Dupuis etal. [21]. Poly(a) + RNA was extracted from 4 x 10 6 isolated male gametes and were processed for cloning into 2ZAP (Stratagene) using the same strategy as for the embryo library. After in vitro encapsidation, approximately 2x 105 recombinant bacteriophages were amplified to produce the working library. For each library, 106 phage were plated and transferred in duplicates onto Hybond-N membranes (Amersham). Filters were prehybridized at 60 °C for 3 h in 6 x SSC (1 x SSC: 150 mM NaC1, 15 mM sodium citrate), 5 x Denhardt's solution, 0.57o SDS and 0.2 mg/ml sheared and denatured herring sperm DNA. Hybridization was performed at 42 °C for 16 h in 6 x SSC, 5 x Denhardt's solution, 50~/o formamide, 0.5~o SDS, 0.2 mg/ml sheared and denatured herring sperm DNA. The "hsf probe' was labelled by random priming (Boehringer-Mannheim) using [32p]dCTP and added to a final concentration of 1.5x 106cpm/ml. Filters were washed for 2 x 30 rain at60 ° C i n 2 x SSC, 0.57o SDS and f o r 2 × 3 0 m i n at 6 0 ° C i n l x SSC, 0.25 Yo SDS. Clone Zmhsfa was isolated from the transition-stage embryo cDNA library and clones Zmhsfb and Zmhsfc from the male gamete library by three rounds of plaque purification. Phagemid s were in vivo excised according to Stratagene's protocol and sequenced as above. Northern blot hybridization analysis Poly(A) + RNA (ca. 0.5 #g) was electrophoresed on denaturing 1.5 Yo agarose/2.2 M formaldehyde gels and blotted onto Hybond-N membranes (Amersham). We have determined by in vitro translation that actin mRNA level does not vary during heat shock at any stage of pollen development (data not shown). Filters were hybridized with an actin probe for comparing poly(A) + RNA amounts in unstressed and heat-shocked samples. We have also noticed by two-dimensional electrophoresis analysis of in vitro translated products that actin mRNA level is much lower in late microspores than in later stages of pollen development but its roughly constant from late-bicellular to mature pollen (data not shown). Thus, the actin control can be used to compare poly(A) + RNA amounts from late-bicellular to mature pollen. All probes were labelled by random priming (Boehringer-Mannheim) using [32p]dCTP. Prehybridization and hybridization were as for library screening. The hsp70 probe corresponds to a conserved 630 bp central part of cytoplasmic hsp70 genes [6]. This probe was obtained by PCR amplification from maize genomic DNA using primers derived from the amino acid sequences GGEDFD and DNQPGV positioned at amino acids 233-238 and 435-440 with respect to the maize gene isolated by Rochester etal. [39]. When using the hsp70 probe, actin probe and phagemid inserts, washes were for 2 x 30 min at 60 °C in 1 x SSC, 0.57o SDS and for 2 x 3 0 m i n at 60°C in 0.1x SSC, 0.25Yo SDS. When using the 101 bp 'hsf probe', washes were performed at 55 °C. 845 RT-PCR analysis ditions. Since the three primer couples amplify fragments which can hybridize to the 101 bp 'hsf probe', amplification products could be unequivocally identified as part of hsf genes by subjecting aliquots of PCR reactions to Southern blot analysis. Hybridization conditions and washes were as for northern analysis. Three hsf gene-specific primer sets were designed from the sequences of three different maize hsf clones, Zmhsfa, b and c (see Fig. 1 for PCR primer positions and sequences, and Results for clone description). Fragments of 291,206 and 108 bp were specifically amplified from clones Zmhsfa, b and c by the primer sets P6-P7, P1-P3 and P1-P5, respectively. RT-PCR experiments were conducted as described for the production of the 101 bp 'hsf probe'. When using the P1-P5 primer set, the amplification efficiency was lower than that observed with the other primer couples and, thus, 1~o of the first PCR reaction was subjected to a second round of PCR under the same con- Results Isolation of microgametophyte populations at five different stages We have isolated homogeneous microgametophyte populations corresponding to five different (a) P6 P7 ,~. Zmhsfa 413 bp % / /'P1 P4~, P3 2048 bp Zmhsfb / /P2 '~ ~' PI P5 - .," 14s0bp Zmhsfc P2 P1 P2 hsfprobe ~"~" 101 bp P1 CCNARRTWYTTYAARCA P2 TCRTTNKCRAAYTSCOA P3 GTATCTGGTAATGGTCC P4 ATCAAGGCTCATAAATc P5 AAAACC'I-I'CATTGGCCC P6 ATGGATGTGCTCCACGACGG P7 CCGACGCCTAATG I i I I 1G 125o bp I (b) Fig. 1. Position and sequence of PCR primers. (a) Schematic representation of Zmhsfa, b, c clones and hsf probe. Position and orientation of PCR primers are indicated by arrows. Regions coding for DNA-binding domain are indicated by hatched rectangle, other coding regions by white rectangle, 5' and 3'-untranslated regions by shaded rectangle, and intron by black rectangle. Note that the portion of Zmhsfc sequence which is underlined has been obtained by 5' RACE-PCR. (b) Sequence of PCR primers. N = (G,A,T, orC);R = ( A o r G ) ; W = (AorT);Y = (CorT);K = ( G o r T ) ; S = ( G o r C ) . 846 stages of pollen development, i.e. late microspores, late-biceUular pollen, early- and midtricellular pollen and mature pollen (Fig. 2). These isolation procedures allowed the purification of 400 000 to 1000 000 gametophytes per tassel. Homogeneity was assessed by checking several cytological criteria that change during pollen development. Hence, starch accumulation as well as vacuole size were monitored according to Alexo ander's procedure [1 ], and the number and condensation state of nuclei were determined by DAPI staining (results not shown). Late microspores, late-bicellular or tricellular pollen never co-purified due to their distinct respective buoyant densities, which depend on the level of starch accumulation. However, a careful selection of spikelets before microspore isolation was necessary to avoid contamination with earlybicellular pollen, the density of which is identical to late microspore density. It should be noted that we never observed contamination by anther tissues. Cell viability was determined using fluorescein diacetate as an indicator of membrane integrity [25] and was 8 5 - 9 0 ~ for late microspores and bicellular pollen and higher than 9 0 ~ for tricellular pollen (results not shown). These isolated fractions correspond to defined developmental stages, display high cellular viability and Fig. 2. Schematicrepresentationof maize pollendevelopment(a) and bright-fieldmicroscopyobservationof isolated microgame- tophytepopulations(b). LM, late microspore;LB, late-biceUularpollen;ET, early-tricellularpollen;MT, mid-tricellularpollen;MP, mature pollen. Cytologicaldetails are indicated: n, nucleus; pw, pollen wall; c, cytoplasm;v, vacuole;vn, vegetativenucleus;gn, generativenucleus; s, starch; scn, spermcell nuclei. Note that the o-ring-shapedstructureat the surfaceof late microspore(LM, b) correspondsto the germinativepore. Bar scale = 100 #m. 847 are devoid of any sporophytic contamination. They are then suitable for molecular investigations of pollen development. Expression ofhsp70 genes during maize pollen development In order to determine the heat-induced and developmentally regulated expression ofhsp70 genes during maize pollen development, the abundance of HSP70 transcripts was investigated at the five stages of pollen development by northern blot analysis (Fig. 3). The probe we used can hybridize to both constitutive and inducible maize cytoplasmic hsp70 genes [6]. Interestingly, in the absence of stress, hsp70 genes were highly expressed at the late-bicellular stage as compared to leaves or other stages of pollen development (Fig. 3a, lanes c). As described earlier for this probe [6], a heat shock-induced accumulation of HSP70 transcripts in leaves (Fig. 3a, lanes L). This increase was also detectable in late microspores but in a significantly weaker manner (Fig. 3a, lanes LM). At the late bicellular stage, the high level of HSP70 m R N A remained constant upon heat shock (Fig. 3a, lanes LB). The signal increase for the heat-shocked early tricellular stage is overestimated due to a larger amount of poly(A) + R N A (Fig. 3a and b). The level of HSP70 m R N A remained constant in case of a heat shock at mild tricellular stage and, for comparable amounts of poly(A) + RNA, no signal was detected in control or heat-stressed mature pollen (Fig. 3a, lanes MP). Thus, in the absence of heat shock, the level of HSP70 transcripts is developmentally regulated in maize microgametophytes, the late-bicellular stage exhibiting a massive accumulation of HSP70 mRNA. In addition, the HSP70 transcript level can be clearly enhanced by a heat shock at the late microspore stage but to a much lesser extent than in vegetative tissues. During the final stages of pollen development, HSP70 m R N A cannot be accumulated in response to a heat shock. Production and characterization of a maize hsf probe Fig. 3. Northern analysis of HSP70 and actin mRNA expressed during pollen development. Poly(A) + RNA (0.5 #g) was extracted from leaves (L), isolated late microspores (LM), late-bicellular (LB), early-triceUular (ET), mid-tricellular (MT) and mature pollen (MP) incubated 2 h either at control (c) or at heat-shock (hs) temperature. (a) Hybridization with the hsp70 probe. A band of 2.5 kb is indicated on the right. (b) Hybridization with the actin probe (see Materials and methods). As we observed an intricate regulation of the expression of hsp70 genes during pollen development, we were interested in analysing the expression of hsf genes during this developmental process. To this end, we produced a probe for detection of maize hsf genes. All known hsf genes exhibit remarkable homology in the region encoding their D N A binding domains ([36], Fig. 4). Two primers, P 1 and P2 (Fig. 1 and Fig. 4), were designed in order to amplify the most conserved part of this region. A 101 bp fragment was amplified by P C R from maize unstressed leafcDNA. Its deduced amino acid sequence is identical to the residues present at position 82-115 of the predicted protein sequence of tomato HSF8 [42] and most of these amino acids are conserved in the deduced protein sequences of 14 H S F s 848 A 4 .4 4 44 ~ 4 44 • 4 4 44 VPAFLTKLWTLVSDPDTDAL CWSPSGNSFH V FD O G VPAFLSKLWTLVEETHTNEF TWSQNGQSFL VLD E 0 VPAFLTRLWTLVSDPDTDAL CWSPSGNSFH VFD Q G V P A F L S K L WT L V E E T H T N E F TWSQNGQSFL VLD E Q VSAFLTRLWTLVEDPETDPL CWSPSGNSFH VFD Q G VPAFLBKLWALVGEAPSNQL TWS QN GQS F L V L D E 0 VPGFLAKLWALVEDPQSDDV CWSRNGENFC I LD E Q Dro VPAFLAKLWRLVDDADTNRL CWTKDQQSFV I QN Q A Sac RPAFVNKLWSMLNDDSNTKL QWAEDGKSFI VTN R E VPN R E Klu RPAFVNKLWSMVNDKSNEKF HWSTSGESIV A±hlPPPFLSKTYDMVEDPATDAIVSWSPTNNSF] VWD P P Tom8 PPPFLVKTYDMVDDPSTDKIVSWSPTNNSFV VWD P P Tom24PAPFLLKTYQLVDOAATDDVISWNEIGTTFV VWK T A Tom3OPPPFLSKTYEMVEDSSTDQVISWSTTRNSFI VWD S H Huml Hum2 Moul Mou 2 ChkA ChkB ChkC ~o Zmb ZB¢ Huml Hum 2 Mou I Mou 2 Chk A ChkB ChkC Dros VWD P H VAN Q A I WD S H SPPFLTKTYDMVDDPTTNAVVSWSAANNSFV LPPFLSKTYEMVODPATDAVVAWTPLGTSFV LAPFLTKVYDMVSDPATDAVISWSAGGGSFV * 4**444**444 **4**** **44*4 QFAKEVLPKYFKHNNMASFVRQLNMYGFRKV R F A K E I L P K Y F K H N N MA S F V R O L N MY G F R Q F A K E V L P K Y F K H N N MA S F V R O L N MY G F R R F A K E I L P K Y F K H N N MA S F V R Q L N MY GF R Q F A K E V L P K Y F K H N N MA S F V R Q L N MY G F R RFAKEILPKYFKHNNMASFVRQLNMYGFRKV RFAKELLPKYFKHNNISSFIRQLNMYGFRKV QFAKELLPLNYKHNNMASFIRQLNMYGFHKI Klu LP~YFK SNFASFVRQLNMYGWHKV A th. I E F S R D L YF RQLNTYGFRKV LP ~NN F SSFV Tom8 F KDLLP YF NNFSSFVRQLNTYGFRKV Tom 2~ ~ ; A L P~YF~ NNFSSFVRQLNTYGFRK Tom30 STTL FF SNFSSFIRQLNTYGFRKV P1 Zma IFGTVLLPRYFKHNNFSSFVRQLNTYGFRKV Zmb EFWRDLLPKYFKHNNFSSFVROLNTYGFKKI Zmc AFSARPLPRHFKHNHFTSFIRQLNTYGFHKV A4 * Huml A 4. GGLVKPERDDT-EFQHPCFLRGQEQLLEN Hum2 SGIVKQERDGPVEFQHPYFKQGQDDLLEN Moul GGLVKPERDDT-EFQHPCFLRGQEQLLEN Mou2 SGIIKQERDGPVEFQHPYFKOGQDDLLEN ChkA GGLVKPEKDDT-EFQHPYFIRGQEHLLEN ChkB SGIVKLERDGLVEFOHPYFRQGREDLLEH ChkC -GMITAEKNSVIEFQHPFFKQGNAHLLEN Oros GGL-RFDRDE-IEFSHPFFKRNSPFLLDQ Sac S G S I Q S S S D D K W ~ ~ k EN ~ LL~ Klu SGSMLSNNDSRW EN HLL~K A. th. 1 . . . . . . . . . DRWEF~NN E G F F L G;Q K ~ ; L ~T ~ Tomb DRW~F EG . . . . . . . . . Tom24 DKW NE N~L ~GQKH~LKT Tom30 . . . . . . . . . DRWE~NEG P2 Zmo . . . . . . . . . DRWEFANEEFLRGQRHLLKNIR Zmb . . . . . . . . . EQWEFANDDFIRGQQHRLKNIH Zmc . . . . . . . . . DRWEWANEGFIKGQKHLLKTIK I I K K K K V H I VH I VH l VH I VH I VHV VAL TS I QDV QDV DP DP l V P D P - V V V V D P D P D P - E Q 50 42 50 t¢2 55 56 51 82 2O7 229 85 7t~ q2 6tl tlt~ q2 ,q3 86 D- 77 E Q 86 EE O D E N D N - K 77 91 91 87 118 242 - K - 26q 118 107 75 94 - - 77 75 76 - *A KRKV KRKV KRKV K R K KRK KRK KRK g R K VR Q V V V V I K R a K S R R K S R R K K R RRK ' 119 111 119 111 124 125 120 150 276 298 143 R R R 132 100 122 R R K RRK R K K 102 100 101 Fig. 4. Amino acid sequence alignmentof the DNA binding domains ofZmHSFa, b and c and of 14 HSFs from various organisms. The respective origin of the different HSFs is indicated by the following abbreviation: Hum, human; Mou, mouse; Chk, chicken; Dro, Drosophila; Sac, Saccharomycescerevisae; Klu, Kluyveromyces lactis; A.th., Arabidopsis thaliana; Tom, tomato, Zm, maize. Asterisks indicate identical amino acids and open triangles homologous residues that are conserved among at least 80% of the sequences. Motifs corresponding to P1 and P2 primers are boxed. (Fig. 4). Therefore, this 101 bp sequence was unambiguously identified as part of a hsf gene. Its nucleotide homology with the three tomato corresponding sequences ranges from 59 to 69~o. Thus, it is likely that this fragment hybridizes to most if not all maize hsf genes and is referred to as "hsf probe'. Cloning of three maize hsf cDNAs In order to isolate hsf genes expressed during maize developmental processes, two c D N A libraries constructed from transition-stage embryos or isolated male gametes were screened with the 'hsf probe'. Fourteen positive clones were isolated and sequence analysis indicated that these 849 clones represented three distinct sequences (a, b, c). Clone a was isolated from the transition-stage embryo library and clones b and c from the male gamete library. The sizes of the longest c D N A of each type obtained were 413 bp (a), 2048 bp (b) and 1290 bp (c), respectively. The length of clone c sequence was further brought to 1450 bp by 5' RACE-PCR experiments. When translated, a region of these cDNAs showed high homology to the D N A binding domain of 14 previously characterized HSFs (Fig. 4). The three clones were consequently designated Zmhsfa, b and c and the predicted proteins Z m H S F a , b and c. When compared to H S F s characterized in yeasts and animals, the putative D N A binding domains of the three maize HSFs exhibit a gap of 11 to 12 residues (Fig. 4). This gap seems particular to higher plant H S F s as it is also found in the tomato and Arabidopsis H S F s ([28, 42], see Fig. 4). As expected, sequence comparison (Clustal method) indicates that the D N A binding domains of Z m H S F a , b and c are more related to their tomato counterparts (57.4 to 80.9~o similarity) than to animal and yeast corresponding motifs (37.2 to 52.1~o similarity). It should be noted that the maize motifs are more similar to mammals and avian HSF1 or H S F A (46.8 to 52.1Yo) than to H S F 2 or HSFB (37.2 to 44.7 ~o). The highest identity scores for the D N A binding domain of Z m H S F a , b and c are 80.9~o with tomato HSF8, 70.2~o with tomato HSF8 and 69.1 Yo with tomato HSF30, respectively. On the other hand, phylogenetic analysis (based on Clustal method) revealed that the D N A binding domain of Z m H S F b is more related to that of tomato HSF24. The nucleotide and deduced protein sequences of the longest clone, Zmhsfb, are presented in Fig. 5. A presumptive start codon is present at position 177. It should be noted that no in frame stop codon was found upstream of this A T G codon and thus, we cannot certify that no additional start codon is used in vivo. However, the bases present in 5' and 3' of this putative start codon correspond well to the A T G context consensus sequence (G, A(C,A)(C, G)ATGG(C,A)G described for maize genes [32]. A stop codon and a poly(A) tail are present at position 1644 and 2039, respectively. Thus, it is very likely that Zmhsfb represents a full-length clone. Surprisingly, the open reading frame of the clone Zmhsfb is interrupted at position 381 by a 543 bp region (Fig. 5). Several arguments indicate that this region corresponds to an intron. First, consensus 5' and 3' splice junction sequences [13, 43] were identified and 6 inframe stop codons are present in this 543 bp region (Fig. 5). Second, as shown in Fig. 6a, fragments with the expected size difference were obtained in P C R experiments using genomic D N A or unstressed leaf c D N A as templates, and primers P 1 and P3 flanking the 543 bp region (see Fig. 1 and Fig. 5 for primer sequences and positions). In addition, an intron was reported at the same position in tomato hsJ8 (EMBL accession number X67599) and Arabidopsis Athsfl [28] genes. We have also detected the presence of an additional sequence in the same region of Zmhsfa and Zmhsfc genes by using genomic D N A or c D N A in P C R experiments (results not shown). In addition to the presence of a poly(A) tail, the length of the clone Zmhsfb, minus that of the intron, corresponds to the size estimated from northern experiment (see below). Thus, the clone Zmhsfb could represent a pre-messenger RNA. Indeed, this pre-messenger R N A was detected in stressed leaves by P C R (Fig. 6b) using primers P 1 and P4 (see Fig. 1 and Fig. 5 for primer sequences and positions). As expected, a 550 bp fragment was amplified only when poly(A) ÷ R N A was reverse transcribed indicating that this pre-messenger R N A is present in heat-shocked leaves and that no genomic D N A contamination occured during RT-PCR experiments. The open reading frame determined for the clone Zmhsfb is 927 nucleotides in length and encodes a protein of 308 amino acids with a predicted molecular mass of 35 370 Da. The putative protein contains both the conserved DNAbinding domain and arrays ofhydrophobic amino acid heptad repeats of the leucine zipper type that are characteristic of H S F s (Fig. 5) [36]. The first leucine zipper region is located immediately after the D N A binding domain and the second one, 850 g~cc~cgc~gca~tctgg±~±ctgg~a~tctcc~tc~±ccg~t±ccg~c~g¢ag~a~t±±tcccttcccccg~cca ccaaccaaccaaacccgagcaaccagcct~agccacctcccatgcgcc~cgccgca9cctga~ag~ag~a~ag~ac~c 80 160 99tgg±ctggagagggATGGAQGGCGCGTCCTCGCTQ~GCCCTTCCTGAGCAAGACGTACGAGATGGTGGACGACCCGG 240 M E ~ A S S L ~ P F L S K T Y E M V D D P CCACGGACG~CGTGgTGGCGTG~ACGCCG~TGGGGACCAGCTTCGTCGTCGCGAACCAGGCCGAG~TCT~GAGGGAT~TG 320 A T D A V V A W T P L G T S F V V A N Q A E F W R D L CTCCCCAAGTACTT•AAGCACAACAACTTCTCCAGcTTCGTG•GGCAACTGAACACCTAC•t••••••c••ccgc••••c L P K Y F K H N N F S S F V R Q L N T Y 4~ ~c~±c~tc~±~ttc~tcgcccc~±ctccc±ctct~±cc~cc±g±~±¢t~g~¢c~atc~g~c~±g~c~c~±~ct 480 ~ctg±c~gct~gctg~±~cccc±~tgc~a~c±cc~ca~g~±g~c~tt±g~g~gc~aa~ctt~±ta~cct~-a-~ 560 ~]aacaottg±±~ogggta~cgtc~tcgg~cgaatgaac±g~cc~±g~atagca±agcgtg~caac±~gg~99 840 c~ataac¢±~aaat±aca~t~¢cc~cta±atatc1a~ta±a±a¢~c±caa±~a~ctt~t±~aga~g¢t±~¢cc¢±¢±c~ 720 aaa¢ccaactagog±tag¢gCacc±~±tgacaaa~gaagtgc±gc~aacggtaaacac±g±¢gaat±ggcatc¢¢±g+~ 8~ . . . . . aa~aaa± . . . . ±o9. . . . . . ~aa~gat ±t cgga~gg~t ac~ ~~~' ~ - g ; g c c t t 9a~ggg~caaat ~ 880 ctgtga±gCtgtg±taatcg¢±¢~cc±gc¢±cttc±¢atg~I3GCTTTAAGAAAATTGATCCTGAACAA~GGAGTTT~ 960 ~G F K K [ D P E Q W E F cAAATGATGATTT~ATTAGGGGA~AA~AG~ACCGAcTGAAAAATATACACAGGCGTAA~CCTATATTCAGC~ATTCATCG 1040 A N D D F I R G Q O H R L K N I H R R K J P I F S H S S CATA~TcAGGGTTCTGGA~CATTA~CAGATACCGAAAGGAGGGATTATGAGGAGGAAA~CGAAAGG~TAAGTGTGA~AA1120 H T O G S G P L P D T E R R D Y E E E ~ E R ~ K C D N T~CAG~T~TGA~CTC~GAGCTTGAAAAGAATGCA~AGA~GAAA~TTGTTACAGAGA~A~GAALGcAGGATCTAGAAGA~A AA~TSE@EKNAQKKLVTEKR~QD~ED 12~ A~T~GATCTTTT~GGAGGATCGGCAGAAGAATC~GATGGCGTATG~AGGGATA~GTACAGGCACCAGGATCTTTCTCT 1280 K ~ I F ~ E D R O K N ~ M A Y ~ R D ~ V Q A P ~ S F S AGCTTTGTGCAGCAACCTGATCATCACGGAAAGAAAAG~AGACTACCA~TACCTATCT~TCTCTACCAAGAITCTAATGC 1360 S F V Q Q P D H H G K K R R L P V P I S L Y Q D S N A TAAGGGGAACCAGGTTGTGCATGG~AGCTTCA~CACCAACCCACCAGCTTGCAGGGAATCAT~TGA~AA~ACGGAATCTT1~0 K G N Q V V H G S F I T N P P A C R E S ~ D K T E S ~ATTGAAC~G~G I GAGAA~TTCc~cGGGAAG~GAGTGAAGcGTC I AAT~T~T~TATGA~GA~GGCc~cCCTGGCCTT S~NS~ENF~RE@SEA~NISYDDGLPGL 1520 CATcTG~T~TCGTTATcACAGAGCTCcATTCGTcCGGAGAAA~TGA~CcC~ATGTGCCATCAccTGTCTCAAGAATGCAT 16~ H L L B L S O S S I R P E K V I P M C H H L S Q E C I ACATCTTCG~CTGGTGcGG~AGA~TcGCTCT~TTCcCGCGATT~-~-~c~tcaacta~c±~c~ct~a~ccc¢c~cc± 1680 H L R L V R E I R S L P A I ccc±ca~a±ccaaccct~t~c~attcac9~c~aa~t~¢c~a~¢c~±~±c~atc~9~a~cc~ct~tcaca~aa~ 1760 ct~±ccgacaa~ggaccaacct~cc~a~cc±ccccac~±a~ca~ct~t~aac9~t~c±±c~ca~ca~±tc 18~0 ctcacc~ag~ccc~gtcc±~a~ac~cca~a~ccc~a±ca~a~ag~a~a~a~c~a±~a~ataa~c~a±c~ 1920 ~cga~a~ag~acc~a~a~±t±~g~aa~aa~at~¢~a~ca~at~a~a~a~aa~ct~c~c±ca 2(300 cctcgg~99agaaaacc±gac~gta±~±c~gtagc~gaaaaaaaaao 20~8 Fig. 5. Nucleotide ~ d deduced amino acid sequences of Zmhs~. The 5'- ~ d 3 ' - u n t r ~ s l a t e d re~ons, as we~ as the intron sequence, ~ e in lower case letters. The open reading flame is in upper-case letters. The limits of the putative DNA-bindmg domain ~ e indicated by s q u ~ e brackets. Bases corresponding to 5' ~ d 3' splicing consensus sequences ~ e underlined. Stop codons, including those present ~ t h i n the intron, ~ e boxed. Circles indicate amino acids potenti~ly implicated in leucine zippers. Position ~ d orientation of primers P1, P2, P3 and P4 ~ e indicated by ~rows. near the carboxy terminus of the predicted protein (Fig. 5). Expression of hsf genes in maize leaves We first studied the expression of hsf genes in control and heat-shocked leaves. Northern blot analysis was performed using the 101 bp 'hsf probe' (Fig. 7a). A transcript of ca. 2.3 kb was detected both under control and heat-shock conditions while two other transcripts of 1.6 and 1.5kb were only revealed in heat-shocked samples. To identify these bands, complementary northern blot analysis was performed with the inserts of clones Zmhsfa, b and c (Fig. 7b). Zmhsfa revealed the constitutively expressed hsf 851 Fig. 6. Detection of an intron in Zmhsfb and presence of the pre-messenger RNA in leaves. (a) PCR experiments were conducted using primers P1 and P3, flanking the intron (see Fig. 1 and Fig. 5 for primer positions), and Zmhsfo plasmid (1), genomic DNA (2) and unstressed leaf cDNA (3). (b) PCR experiments were conducted using primers P1 and P4, the latter one being located within the intron (see Fig. 1 and Fig. 5 for primer positions), and: Zrnhsfl9 plasmid (1), heat-shocked leaf cDNA (2), poly(A) ÷ RNA, isolated from heat-shocked leaves without reverse transcription (3). Ten #1 of PCR reactions were analyzed by Southern blot using the conserved 'hsf probe'. Filters were exposed for 30 min. transcript (2.3 kb) whereas Zmhsfo and Zmhsfc hybridized with the inducible 1.6 kb and 1.5 kb transcripts, respectively. The expression of the three hsf genes was further investigated by RT-PCR using gene-specific primer sets (Fig. 7c). As expected from the previous northern experiments, Zmhsfa mRNA was detected in roughly equal amounts both in control and heat-shocked leaves. Interestingly, Zmhsfb and Zmhsfc mRNAs were revealed in unstressed leaves and the signals were slightly enhanced upon heat shock (Fig. 7c). This signal increase was reproducibly observed in several experiments, the constant signal for Zmhsfa gene being used as internal standard. These results clearly show that at least three hsf genes are expressed in maize leaves. The level of Zmhsfa mRNA is independent of temperature whereas the levels of Zmhsfb and Zmhsfc RNA are increased upon heat shock. Expression of HSF mRNA duringpollen development The developmentally regulated and heat-induced expression of hsf genes was then analysed at five Fig. 7. H S F mRNA expression in unstressed and heat-shocked maize leaves. Poly(A) ÷ RNA was extracted from leaves of 10-day old seedfings incubated either 2 h at control (C) or at heat shock (HS) temperature. (a) Northern analysis was performed using 0.5 #g of poly(A) ÷ RNA and the 101 bp 'hsf probe'. Transcript sizes are indicated in kb on the right. (b) Northern analysis was performed using the three clones Zmhsfa, b and c as probes. (c) RT-PCR analysis was conducted using primers P6 and P7 specific for Zmhsfa (lanes 1, band labelled with a), primers P1 and P3 for Zmhsfo (lanes 2, band labelled with b) and primers P1 and P5 for Zmhsfc (lanes 3, band labelled, with c). In the latter ease, 1 Yo of the reaction was subjected to a second round of PCR. Ten #1 of PCR reactions were subjected to Southern analysis using the conserved 'hsf probe'. Filters were exposed for 30 rain. Note that in lane HS, 2, a signal which size corresponds to the unspliced form of Zmhsfb mRNA is indicated by an arrow. 852 Fig. 8. Northern analysis of HSF mRNA expression during maize pollen development. Late microspores (LM), latebicellular pollen (LB), early- and mid-tricellular pollen (ET and MT, respectively) and mature pollen (MP) were incubated either 2 h at control (C) or at heat-shock (HS) temperature. About 0.5/2g poly(A) ÷ RNA were subjected to northern analysis using the conserved 101 bp 'hsf probe' and the filter was exposed for 2 weeks. A band of 2.3 kb is indicated on the right. A faint band present in LM lanes may not be clearly visible on the photograph. Lanes LM to MT are from the same filter as in Fig. 3. See Fig. 3b for control of poly(A) + RNA amounts after hybridization with an actin probe for these lanes. stages of maize pollen development, i.e. late microspore, late-bicellular pollen, early- and midtricellular pollen and mature pollen. The presence of H S F m R N A was first investigated by northern blot analysis using the 'hsf probe' which recognises the three maize hsf genes (Fig. 8). A faint band corresponding to Zmhsfa m R N A (2.3 kb transcripts) was detected at all stages of pollen development. For each stage, the level of Zmhsfa transcripts was not modified by heat shock (Fig. 8), as it was observed for vegetative tissues. Interestingly, the expression of the two heatinducible hsf genes could not be detected after heat shock at any stage of pollen development. Because this observation, as well as the weak signal detected in late microspores, could be due to a lack of sensitivity of our northern blot analysis, the expression of the three hsf genes was also investigated by RT-PCR using gene-specific pairs of primers (Fig. 9). Zmhsfa m R N A was detected throughout microgametophyte developmental process (Fig. 9, lanes 1), confirming the result of the northern blot analysis. In addition, Zmhsfb and Zmhsfc m R N A s were detected at the different stages of immature pollen. However, in contrast with leaves, their signals were not increased upon heat shock (Fig. 9, lanes 2 and 3). In mature pollen, the signal corresponding to Zmhsfb was no longer detectable. This result was confirmed by four independant preparations of cDNA. However, this phenomenon is probably not related to the loss of hsp gene inducibility since Zmhsfb transcripts can be detected at the mid tricellular stage, a stage at which no H S P m R N A synthesis can be induced. In conclusion, both northern and RT-PCR analyses show that Zmhsfa m R N A is constitutively present throughout pollen development including stages unable to trigger a heat shock response, as mid-tricellular or mature pollen. In Fig. 9. RT-PCR analysis o f H S F mRNA expression during pollen development. Late microspores (lm), late-bicellular pollen (LB), early- and mid-tricellular pollen (ET and MT, respectively) and mature pollen (MP) were incubated either 2 h at control (C) or at heat-shock (HS) temperature. PCR experiments were conducted using primers P6 and P7 specific for Zmhsfa (lanes 1, bands labelled with a), primers P1 and P3 for Zmhsfo (lanes 2, bands labelled with b) and primers P1 and P5 for Zmhsfc (lanes 3, bands labelled with c). When using this latter couple of primers, a second round of PCR was performed. Ten #1 of PCR reactions were subjected to Southern analysis using the conserved 101 bp 'hsf probe'. Filters were exposed for 30 min. 853 contrast to leaves, the basal transcript levels of Zmhsfb and Zmhsfc genes are not significantly enhanced upon heat shock, at any stage of pollen development. These results provide evidence that the loss of hsp gene activation at the final stages of pollen development is not due to a modification of expression of these three hsf genes at the mRNA level. Discussion Expression of three hsf genes in maize vegetative tissues We report here the expression of three hsf genes in maize leaves; Zmhsfa gene is constitutively expressed whereas the expression of Zmhsfb and Zmhsfc genes is strongly enhanced upon heat shock. Similar features have been previously described for tomato hsf genes [42]. From the phylogenetic analysis of the DNA-binding domain sequences, the sizes of the transcripts and the type of expression, it is likely that Zmhsfa is the maize counterpart of tomato hsJ8 gene and, Zmhsfb and Zmhsfc that of hsf24 and hsf30 genes, respectively. Recently, a hsf gene, A thsfl, has been characterized in A rabidopsis [ 28 ]. Its expression is enhanced upon heat shock. Hence, heatinduced expression of hsf genes has been characterized to date in three plant species and seems to be specific to higher plants. The three tomato HSFs are able, independently, to stimulate upon heat shock the transcription of a reporter gene fused with various promoters conferring heat inducibility [48]. However, the activation by each HSF is dependent on the position and numbers of target sequences, i.e. HSEs, in the promoter of the reporter gene but also on the stress regime [48]. Hence, the different plant HSFs could be required to modulate the expression of hsp genes depending on the stress conditions. Developmentally regulated expression ofhsp70 and hsf genes during pollen development In plants, several reports have focused on the expression of hspl8 and hsp90 gene during devel- opment of the whole anther, i.e. a combination of sporophytic and gametophytic tissues [4, 9, 10, 19, 33]. In the present work, the use of isolated microgametophyte populations allowed us to characterize the expression of hsp70 genes at the mRNA level during pollen development without taking into account the sporophytic tissues of the anthers. Using an in situ hybridization approach Duck and Folk [20] have also detected recently HSP70 mRNA at non-heat shock temperatures in developing tomato pollen (10 mm buds). Similarly, the expression ofhsp70 genes has been characterized during embryogenesis and gametogenesis in amphibians and mammals [8, 14, 34, 53] and during seed development in plants [ 18, 52]. Hence, members of hsp70 family are expressed in plant and animal developmental processes but their precise biological role is still largely unknown. In mammals and chicken, the developmental regulation ofhsp genes appears to be mediated by specific HSFs [36]. Indeed, the DNA binding activity of mammalian H S F I and avian HSFA is activated by various stresses, whereas HSF2 is activated during hemin-induced differentiation of erythroleukemia cells [44]. Strikingly, a HSEbinding activity was detected in non-stressed mouse embryonal carcinoma cells but this activity disappeared after in vitro differentiation into fibroblasts [34]. Also, a H SF2 activity is found in the absence of heat stress in the mouse testis [41] and in mouse blastocysts [ 35 ]. Thus, H S Fs might be implicated in hsp gene regulation during developmental processes. Further investigations are needed to determine whether the stress-independent regulation of hsp70 genes during maize pollen development is mediated by the H S F encoded by the constitutively expressed Zmhsfa gene. We did not identify a fourth kind of HSF mRNA (in addition to the three HSF mRNA species that we have detected in leaves) by screening two cDNA libraries of developing embryos and mole gametes or by northern blot analysis of developing pollen. However, in the latter case, a fourth type of HSF mRNA would have been revealed only if its size was different and if it was sufficiently expressed to be detected by this technique. 854 Expression of hsp70 and hsf genes during pollen development under heat stress conditions In the present report, we show that HSP70 transcripts do not accumulate significantly in response to a heat shock after the late microspore stage. We have also observed by RT-PCR and in vitro translation that hsp18 genes can be heat-induced only until the early-tricellular stage (unpublished results). Thus, the gradual loss of HSP synthesis in response to a heat shock, that was first reported by Frova et aL [23] during pollen development, seems to be due to a stage-dependent defect in accumulating HSP mRNA. Our results show that the defect in hsp gene expression during the final stages of maize pollen development is not due to a lack of Zmhsfa mRNA. Interestingly, a HSE-binding activity can be induced by a heat shock in Xenopus unfertilized eggs and cleavage stage embryos, stages at which hsp gene expression is inhibited [37]. A similar phenomenon might occur at the final stages of pollen development. In contrast to leaves, the low basal expression of Zmhsfb and Zmhsfc genes is not significantly enhanced under heat stress during pollen development, although we have shown that hsp genes can be expressed in immature pollen. Therefore, it appears that the immediate heat-induced hsp gene transcription observed in immature pollen can occur without an increase of expression of the two heat inducible hsf genes, Zmhsfb and Zmhsfc. Nevertheless, we cannot exclude a very effective translation of the minute amounts of Zmhsfb and Zmhsfc mRNA that we have detected and which could lead to the accumulation of the corresponding proteins. Indeed, such a regulation was described in carrot globular embryo s for H S P mRNAs which are very efficiently translated upon heat shock [3]. On the other hand, the heatinduced accumulation of H S P transcripts is much weaker during pollen development than in vegetative tissues. This phenomenon could be correlated to the lack of heat inducible hsf gene expression in immature pollen. Promoter accessibility is likely to play an important role in regulating hsp gene expression in eukaryotes. Both human and Drosophila HSFs fail to bind in vitro to HSEs packaged in nucleosomes [7, 47] and specific factors, the GAGA factors, have been shown to disrupt nucleosomes present on hsp70 promoters [50]. Thus, it could be interesting to investigate hsp gene promoter accessibility in developing pollen and in other tissues which exhibit a complete heat shock response. This might provide further insight into the gradual loss ofhsp gene activation during pollen development. In conclusion, we show that the defect in accumulating HSP70 transcripts during pollen development is not correlated with a modification of Zmhsfa, b or c gene expression. Our results suggest that the weak, immediate accumulation of HSP mRNA in response to a heat shock could be mediated in immature pollen by the HSF encoded by the constitutively expressed hsf gene, Zmhsfa. Acknowledgements We thank Elizabeth E. M. Bates for providing the hsp70 probe, Michel Beckert (Institut National de la Recherche Agronomique, Clermond-Ferrand, France) for DH5 x DH7 maize seeds, Richard Blanc for growing the maize plants, Pierre Audenis for photographic work and Fran~oise Mon6ger and Peter Rogowsky for critical reading of the manuscript. This work was supported by the Association Grnrrale des Producteurs de Ma'is (Pau, France) and by a Human Frontier Science Program grant to C. D. References 1. Alexander MP: Differential staining of aborted and nonaborted pollen. Stain Technol 44:117-122 (1969). 2. Ang D, Liberek K, Skowyra D, Zylicz M, Georgopoulos C: Biological role and regulation of the universally conserved heat shock proteins. J Biol Chem 266: 2423324236 (1991). 3. Apuya NR, Zimmerman JL: Heat shock gene expression is controlled primarily at the translational level in carrot cells and somatic embryos. Plant Cell 4:657-665 (1992). 855 4. Atkinson BG, Raizada M, Bouchard RA, Frappier JRH, Walden DB: The independent, stage-specific expression of the 18-kDa heat shock protein genes during microsporogenesis in Zea mays L. Devel Genet 14:15-26 (1993). 5. Barloy D, Denis L, Beckert M: Comparison of the aptitude for anther culture in some androgenetic doubled haploid maize lines. Maydica 34:303-308 (1989). 6. Bates EEM, Vergne P, Dumas C: Analysis of the hsp70 gene family in Zea mays. Plant Mol Biol 25:909-916 (1994). 7. Becker PB, Rabindran SK, Wu C: Heat shock-regulated transcription in vitro from a reconstituted chromatin template. Proc Natl Acad Sci USA 88:4109-4113 (1991). 8. Billoud B, Rodriguez-Martin ML, Berard L, Moreau N, Angelier N: Constitutive expression of a somatic heatinducible hsp70 gene during amphibian oogenesis. Development 119:921-932 (1993). 9. Bouchard RA: Characterization of expressed meiotic prophase repeat transcript clones of Lilium: meiosisspecific expression, relatedness, and affinities to small heat shock protein genes. Genome 33:68-79 (1990). 10. Bouchard RA, Frappier JRH, Liu L, Raizada M, Atkinson BG, Walden DB: Developmentally-modulated expression of transcripts from stress-inducible gene families during microsporogenesis and gametophyte development in Zea mays L. Maydica 38:135-144 (1993). 11. Breton C, Chaboud A, Matthys-Rochon E, Bates EEM, Cock JM, Fromm H, Dumas C: PCR-generated cDNA library of transition-stage embryos: cloning and expression of calmodulin genes during early embryogenesis. Plant Mol Biol 27:105-113 (1995). 12. Brewbaker JL, Kwack BH: The essential role of calcium ions in pollen germination and pollen tube growth. Am J Bot 50:859-865 (1963). 13. Brown JWS: A catalogue of splice junctions and putative branchpoint sequences from plant introns. Nucl Acids Res 14:9549-9559 (1986). 14. Christians E, Campion E, Thompson EM, Renard J-P: Expression of the H S P 70.1 gene, a landmark of early zygotic activity in the mouse embryo, is restricted to the first burst of transcription. Development 121:113-122 (1995). 15. Coca MA, Almoguera C, Jordano J: Expression of sunflower low-molecular-weight heat-shock proteins during embryogenesis and persistence after germination: localization and possible functional implications. Plant Mol Biol 25:479-492 (1994). 16. Cooper P, Ho TD, Hauptmann RM: Tissue specificity of the heat-shock response in maize. Plant Physiol 75: 431441 (1984). 17. DeRocher AE, Vierling E: Developmental control of smaU heat shock protein expression during pea seed maturation. Plant J 5:93-102 (1994). 18. DeRocherA, VierlingE: Cytoplasmic HSP70 homologues of pea: differential expression in vegetative and embryonic organs. Plant Mol Biol 27:441-456 (1995). 19. Dietrich PS, Bouchard RA, Casey ES, Sinibaldi RM: Isolation and characterization of a small heat shock protein gene from maize. Plant Physio196: 1268-1276 (1991). 20. Duck NB, Folk WR: Hsp70 heat shock protein cognate is expressed and stored in developing tomato pollen. Plant Mol Biol 26:1031-1039 (1994). 21. Dupuis I, Roeckel P, Matthys-Rochon E, Dumas C: Procedure to isolate viable sperm cells from corn (Zea mays L.) pollen grains. Plant Physiol 85:876-878 (1987). 22. Dupuis I, Dumas C: Influence of temperature stress on in vitro fertilization and heat shock protein synthesis in maize (Zea rnays L.) reproductive tissues. Plant Physiol 94: 665-670 (1990). 23. Frova C, Taramino G, Binelli G: Heat-shock proteins during pollen development in maize. Devel Genet 10: 324-332 (1989). 24. Gaillard A, Vergne P, Beckert M: Optimization of maize microspore isolation and culture conditions for reliable plant regeneration. Plant Cell Rep 10:55-58 (1991). 25. Heslop-Harrisson J, Heslop-Harrisson Y: Evaluation of pollen viability by enzymatically induced fluorescence: intracellular hydrolysis of fluorescein diacetate. Stain Technol 45:115-120 (1970). 26. Hightower L, Nover L: Heat Shock and Development. Springer-Verlag, Berlin/Heidelberg (1991). 27. Hopf N, Plesofsky-Vig N, Brambl R: The heat shock response of pollen and other tissues of maize. Plant Mol Biol 19:623-630 (1992). 28. Httbel A, SchOffl F: Arabidopsis heat shock factor: isolation and characterization of the gene and the recombinant protein. Plant Mol Biol 26:353-362 (1994). 29. Kerhoas C, Gay G, Dumas C: A multidisciplinary approach to the study of the plasma membrane of Zea mays pollen during controlled dehydration. Planta 171:1-10 (1987). 30. Koning AJ, Rose R, Comai L: Developmental expression of tomato heat-shock cognate protein 80. Plant Physiol 100:801-811 (1992). 31. Lis J, Wu C: Protein traffic on the heat shock promoter: parking, stalling, and trucking along. Cell 74:1-4 (1993). 32. Luehrsen KR, Walbot V: The impact of AUG start codon context on maize gene expression in vivo. Plant Cell Rep 13:454-458 (1994). 33. Marrs KA, Casey ES, Capitant SA, Bouchard RA, Dietrich PS, Mettler IJ, Sinibaldi RM: Characterization of 2 maize HSP90 heat shock protein genes: expression during heat shock, embryogenesis, and pollen development. Devel Genet 14:27-41 (1993). 34. Mezger V, Legagneux V, Babinet C, Morange M, Bensaude O: Heat shock protein synthesis in preimplantation mouse embryo and embryonal carcinoma cells. In: Hightower L, Nover L (eds) Heat Shock and Development, pp. 153-166. Springer-Verlag, Berlin/Heidelberg (1991). 35. Mezger V, Rallu M, Morimoto RI, Morange M, Renard 856 36. 37. 38. 39. 40. 41. 42. 43. 44. J-P: Heat shock factor 2-like activity in mouse blastocysts. Devel Biol 166:819-822 (1994). Morimoto RI: Cells in stress: transcriptional activation of heat shock genes. Science 259:1409-1410 (1993). Ovsenek N, Heikkila JJ: DNA sequence-specific binding activity of the heat shock transcription factor is heatinducible before the midblastula transition of early Xenopus development. Development 110:427-433 (1990). Pr~ndl R, Kloske E, Sch6ffl F: Developmental regulation and tissue-specific differences of heat shock gene expression in transgenic tobacco and Arabidopsis plants. Plant Mol Biol 28:73-82 (1995). Rochester DE, Wirier JA, Shah DM: The structure and expression of maize genes encoding the major heat shock protein, hsp70. EMBO J 5:451-458 (1986). Rogers SO, Bendich AJ: Extraction of DNA from plant tissues. In: Gelvin SB, Schilperoort RA, Verma DPS (eds) Plant Molecular Biology Manual, pp. A6/l-A6/11. Kluwer Academic Publishers, Dordrecht (1988). Sarge KD, Park-Sarge OK, Kirby JD, Mayo KE, Morimoto RI: Expression of heat-shock factor 2 in mouse testis: potential role as a regulator of heat-shock protein gene expression during spermatogenesis. Biol Reprod 50: 1334-1343 (1994). Scharf KD, Rose S, Zott W, SchOffF, Nover L: Three tomato genes code for heat stress transcription factors with a region of remarkable homology to the DNAbinding domain of the yeast HSF. EMBO J 9:4495-4501 (1990). Sinibaldi RM, Mettle IJ: Intron splicing and intronmediated enhanced expression in monocots. Prog Nucl Acid Res Mol Biol 42:229-257 (1992). Sistonen L, Sarge KD, Phillips B, Abravaya K, Morimoto RI: Activation of heat shock factor 2 during hemin-induced differentiation of human erythroleukemia cells. Mol Cell Biol 12:4104-4111 (1992). 45. Sorger PK: Heat shock factor and the heat shock response. Cell 65:363-366 (1991). 46. SpenaA, Schell J: The expression ofaheat-induciblechimeric gene in transgenic tobacco plants. Mol Gen Genet 206:436-440 (1987). 47. Taylor ICA, Workman JL, Schuetz TJ, Kingston RE: Facilitated binding of GAL4 and heat shock factor to nucleosomal templates: differential functions of DNAbinding domains. Genes Devel 5:1285-1298 (1991). 48. TreuterE, NoverL, Ohme K, Scharf KD: Promoter specificity and deletion analysis of 3 heat stress transcription factors of tomato. Mol Gen Genet 240:113-125 (1993). 49. Tsukaya H, Takahashi T, Naito S, KomedaY: Floral organ-specific and constitutive expression of an Arabidopsis thaliana heat-shock HSP18.2::GUS fusion gene is retained even after homeotic conversion of flowers by mutation. Mol Gen Genet 237:26-32 (1993). 50. Tsukiyama T, Becker PB, Wu C: ATP-dependent nucleosome disruption at a heat-shock promoter mediated by binding of GAGA transcription factor. Nature 367: 525532 (1994). 51. Vierling E: The roles of heat shock proteins in plants. Annu Rev Plant Physiol Plant Mol Biol 42:579-620 (1991). 52. Winter J, Sinibaldi R: The expression of heat shock protein and cognate genes during plant development. In: Hightower L, Nover L (eds) Heat Shock and Development, pp. 85-105. Springer-Verlag, Berlin/Heidelberg (1991). 53. Wolgemuth DJ, Gruppi CM: Heat shock gene expression during mammalian gametogenesis and early embryogenesis. In: Hightower L, Nover L (eds) Heat Shock and Development, pp. 138-152. Springer-Verlag, Berlin/ Heidelberg (1991).