J Sci Food Agric 1998, 76, 505È514 Eþect of Down-Regulation of Cinnamyl Alcohol Dehydrogenase on Cell Wall Composition and on Degradability of Tobacco Stems Marie Andre e Bernard Vailhe ,1 Jean Michel Besle,1* Marie Paule Maillot,1 Agnès Cornu,1 Claire Halpin2 and Mary Knight2 1 SRNH, INRA de Clermont Ferrand-Theix, 63122 St Genès Champanelle, France 2 Zeneca Seeds, Jealotts Hill Research Station, Bracknell, Berkshire, RG12 6EY, UK (Received 27 June 1997 ; revised version received 15 July 1997 ; accepted 7 August 1997) Abstract : The e†ect of down-regulation of tobacco cinnamyl alcohol dehydrogenase (CAD) on cell wall composition and degradability has been assessed. CAD activity was only 20, 16, 14 and 7%, relative to the controls, in four populations of plants (designated 40-1, 40-2, 48 and 50, respectively) transformed with CAD antisense mRNA. Cell wall residues of stem samples were analysed for polysaccharide composition, gravimetric and acetyl bromide lignins and lignin nitrobenzene oxidation products. In situ disappearance and cellulase solubility of both initial dry matter and CWR were determined. The populations of plants with depressed CAD activity showed no change in lignin content but some consistent changes in cell wall composition and digestibility were identiÐed. The syringyl content of lignins decreased and the syringaldehyde to vanillin ratio (S/V) was consequently reduced. Dry matter degradability, as measured by both methods, was signiÐcantly improved in all CAD-depressed samples except for population 40-1, which was the least CAD-depressed. Increased in situ disappearance of cell wall (ISCWD) was found in all plants exhibiting more than 80% CAD downregulation and was maximal (7 percentage units) in population 50 which had the greatest CAD depression. The rates of ISCWD increased slightly in some populations (40-2 and 50). The relationship between S/V and ISCWD was signiÐcant (R \ [0É68) only in the samples from a selected population of mature, most depleted plants. Other modiÐcations may therefore also contribute to the improvement in degradability. However the changes in lignin composition that were observed in CAD-depressed tobacco are largely similar to those seen in some maize and sorghum mutants with altered ligniÐcation and improved digestibility. These data therefore suggest that depressing CAD activity may be an e†ective method for improving digestibility in forage crops. ( 1998 SCI. J Sci Food Agric 76, 505È514 (1998) Key words : down-regulation ; cinnamyl alcohol dehydrogenase ; CAD ; cell wall ; lignin ; degradability ; transgenic ; genetic engineering ; tobacco INTRODUCTION 1980). Several studies suggest that lignins play a role not only through their concentration but also through their structural characteristics (Jung and Deetz 1993 ; Besle et al 1994). The concentration and composition of lignins varies in di†erent plants (Monties 1986) and it may be possible to modify lignins without interfering with normal plant development. One way of modifying Lignins are the most important compositional factor a†ecting the utilisation of forages by ruminants (Jarrige * To whom correspondence should be addressed. Contract/grant sponsor : EUÈECLAIR programme. Contract/grant number : AGRE 0021 OPLIGE. 505 ( 1998 SCI. J Sci Food Agric 0022È5142/98/$17.50. Printed in Great Britain M A Bernard V ailhe et al 506 ligniÐcation is to inhibit enzymes of the lignin biosynthetic pathway. Cinnamyl alcohol dehydrogenase (CAD), which catalyses the last step of the synthesis of lignin monomers, has been successfully inhibited by chemical inhibitors in poplar tissues (Grand et al 1985) and in wheat (Moerschbacher et al 1990) but the e†ects on lignin composition and digestibility of the cell walls were not determined. Mutants with altered ligniÐcation have been identiÐed in some graminaceous species. A maize mutant with a reddish brown pigmentation of the leaf midrib was discovered in 1924 (Jorgenson 1931). It was classiÐed as bm1 (Kuc and Nelson 1964) and is now considered to have altered CAD activity (Halpin et al 1995). Several studies have shown that the digestibility of this mutant is greater than that of the normal counterpart both in vitro and in vivo (Barrière and Argillier 1993). Its rate of digestibility is either the same or slightly greater than that of normal plants (Cymbaluk et al 1973). Similarly, it has been shown that the bmr6 mutant of sorghum has greater in vitro digestibility (Hanna et al 1981) and lower CAD activity than its normal counterpart (Bucholtz et al 1980) but, in this case, a slightly lower level of O-methyl transferase activity was also detected (Pillonel et al 1991). The improvement in digestibility was correlated with a lower lignin content and with differences in lignin structure, particularly a lower syringyl content in lignins for both mutants (Gee et al 1968 ; Pillonel et al 1991). These Ðndings prompted attempts to carry out downregulation of CAD by genetic engineering. The Ðrst successful attempts to regulate CAD activity were made by transforming tobacco (Knight et al 1992 ; Halpin et al 1994) and poplar (Baucher et al 1996) with antisense genes for CAD. In these plants, lignin composition was modiÐed in ways similar to that found in CADdepressed maize and sorghum mutants. There was a decrease in syringyl units and an increase in the amount of lignin that could be extracted with alkali. However, unlike the mutants, the transgenic plants had normal levels of lignin. The aim of the present study was to assess whether cell wall modiÐcation in the CAD depressed tobacco of Halpin et al (1994) a†ected degradability. MATERIALS AND METHODS Plants Tobacco plants (Nicotiana tabacum var Samsun), transformed via Agrobacterium with a CAD antisense construct, were produced as described in Halpin et al (1994). Three independent primary transformants (plants 40, 48 and 50) that have a single transgenic locus and reduced levels of CAD activity were selected. These were backcrossed and selfed to produce F1 (backcrossed) and F1 (selfed) seed. Progeny (120 plants) of the backcross of plants 40 and 50 were grown from seed in a greenhouse in the summer of 1992. The genotype of each plant was determined by PCR of the transgene ; segregation of the transgene was as expected (50% hemizygous antisense plants and 50% azygous control plants). In lines 40 and 50, CAD activity of antisense plants was only 20 and 7%, respectively, of that of the corresponding controls. Ten antisense and 10 control plants for each line were harvested after 16 weeks and stems were fresh-frozen. These samples were designated 50 (from line 50) or 40-1 (from line 40). Homozygous line 40 plants were produced by selÐng F1 (selfed) plants and studying the segregation of kanamycin resistance in F2 seedling progeny. Twenty homozygous antisense and 20 control F2 plants were identiÐed from F2 seeds. These were randomly positioned in a greenhouse and grown during the spring of 1993. The CAD activity in antisense plants was 16% that of azygous controls. At 8 and 14 weeks, 5 antisense and 5 control plants for each stage were harvested and stems were fresh-frozen. Samples from 8-week-old and from 14-week-old plants were designated 40-2-8 and 40-2-14, respectively. Homozygous line 48 antisense plants were produced as for line 40. F2 plants were grown in a greenhouse during the winter of 1993. Four antisense plants, with 14% of the CAD activity of the controls, and 4 control plants were harvested after 16 weeks and the stems were fresh-frozen. These samples were designated 48. The control and antisense stems were freeze-dried. Stems of lines 40-1 and 50 were separated into internodes from the top (5È6 internodes), the middle (3È4 internodes), the base (3È4 internodes) and a remaining fraction. The whole stems of lines 40-2 and 48 were studied. All the fractions or whole stems were ground with a blade mill (Brabender 880803, Duisburg) through a 2 mm screen and sieved through a 1 mm sieve. The retained fractions were repeatedly ground to pass through the sieve. The Ðne particles passing through a 100 lm sieve were discarded to minimise physical losses from the bags used for the in situ degradability determinations. The proportions of these Ðne particles were similar for the controls and antisense samples (about 4È7% DM for the di†erent lines). The particles originated mostly from the pith and had a very low acetyl bromide lignin content. For lines 40-1 and 50, whole stem samples were reconstituted from all ground fractions. Chemical analyses Analyses were performed on the cell wall residues (CWRs) prepared in duplicate after sequential washing Cell wall composition and degradability of tobacco stems 507 of the initial sample with water (three times for 15 min each at 40¡C) and Soxhlet reÑuxing with ethanol 95% (v/v) and toluene/ethanol (2 : 1, v/v) (Jarrige 1961). Lignin content was determined in duplicate by the gravimetric method (JL) of Jarrige (1961) and in triplicate by the acetyl bromide (ABL) spectrophotometric method of Iiyama and Wallis (1990) with ferulic acid as standard. Uncondensed units in lignin were determined in duplicate as aldehydes released by the nitrobenzene oxidation method of Venverloo (1971), modiÐed according to Monties (pers comm). NaOH (2 M, 5 ml) and nitrobenzene (0É5 ml) were added to the sample (20 mg) ; hydrolysis was performed at 160¡C for 3 h in an oil bath ; the sample was extracted into dichloromethane and, after vacuum removal of the solvent by evaporation, dissolved in 2 ml methanol. The main products obtained (p-OH benzaldehyde, vanillin \ V, syringaldehyde \ S) were designated as NBO. The methanolic extracts were analysed by high-performance liquid chromatography (HPLC) with a reverse-phase column (Lichrosorb RP-18, 250 ] 4É6 mm id) as described by Mosoni et al (1994). The CWRs were saponiÐed in duplicate according to Scalbert et al (1985) with the following modiÐcations. A sample (1 g) was extracted with 1 M NaOH (30 ml) under N at 35¡C for 2 h, then centrifuged. The pellet 2 was treated again and then water-washed through a sintered crucible (porosity 2). The supernatants and Ðltrates were pooled, acidiÐed with 6 M HCl to pH 2, left for 16 h at 4¡C and centrifuged. The pellets were solubilised in NaOH, dialysed (membrane cuto† 6000È 8000 MW) and freeze-dried to yield the alkali soluble cell wall polymers (ASCWP). EsteriÐed phenolic acids were extracted from the acid supernatants with ethyl acetate and further analysed by HPLC as for the products of nitrobenzene oxidation. Uronic acids were determined (triplicates) by the method of Blumenkrantz and Asboe-Hansen (1973) in hydrolysates prepared according to Englyst et al (1982). Non-cellulose polysaccharides and cellulose were sequentially hydrolysed (duplicates) according to Jarrige (1961) and determined gravimetrically (all intermediate dryings were performed at 60¡C). In addition, for lines 40-1 and 50, the hydrolysates were analysed colorimetrically for reducing sugars using xylose and glucose respectively as standards (Besle et al 1981). The dry weight was measured after 24 h at 60¡C in a conventional oven and the dry matter (DM) after a further 24 h at 103¡C. 40-2, and CWRs for line 48. Losses of water-soluble and Ðne particulate matter were determined by placing duplicate samples in bags in a vessel containing water vigorously shaken for 4 h at 40¡C, and then by Ðltering the extracts on sintered glass (porosity 2) (MichaletDoreau 1990). After incubation (2, 6, 12, 24, 48 and 72 h), the bags were frozen to detach part of the associated microorganisms, thawed and then thoroughly washed with tap water (at about 30¡C) in an electric washing machine and dried at 60¡C. The in situ dry matter disappearance (ISDMD) was calculated on the basis of dry weight at 60¡C taking into account the Ðne particle losses. Since all the soluble cell material would have been extracted by water during the preparation of CWRs, the kinetics of CWR disappearance (ISCWD) were calculated on the assumption that cellular material was totally solubilised during the Ðrst 2 h of rumen incubation. Cellulase digestibility determinations were carried out in duplicate according to the method of Rexen (1977) modiÐed by Cordesse (1982). A sample (500 mg) was incubated with 30 ml of acetate bu†er (0É05 M, pH 4É6) containing 10 g litre~1 Onozuka R10 cell wall degrading complex and 0É3 g litre~1 NaN , and incubated for 3 24 h at 40¡C with mechanical shaking. The residue was then Ðltered on a sintered crucible (porosity 2) and dried for 48 h at 60¡C. The enzyme solution was freshly prepared and previously Ðltered through a sintered crucible (porosity 2). Evaluation of digestibility In situ disappearance was measured with nylon bags (50 lm porosity) placed in triplicate in cannulated sheep which were fed lucerne hay (Demarquilly and Chenost 1969). Initial DM was used for lines 40-1, 50 and Experimental design The kinetics of in situ disappearance were performed on whole stems, and 48 h-in situ disappearances and cellulase digestibility were determined on all samples (stems and stem fractions). The in situ measurements were made on sheep kept under controlled and conÐned conditions, and receiving the same lucerne forage for all the trials. In each in situ trial, a standard wheat straw sample was incubated in triplicate for 48 h and no signiÐcant di†erence in the in situ dry matter disappearance occurred between trials (mean 32É3 ^ 1É3%). Statistical analysis The kinetics of in situ dry matter and cell wall disappearance were Ðtted (NLIN procedures of SAS (1989)) to an exponential model without (^rskov and McDonald 1979) or with (Dhanoa 1988) lag time. The GLM procedure of SAS (1989) was used. The data of CWR content and composition, kinetic parameters, 48 h-in situ disappearances and cellulase solubilities were analysed in a complete block design with two or three replications and subjected to one-way analysis of variance. The data from line 40-2 degradability were subjected to two-way analysis of variance M A Bernard V ailhe et al 508 with the e†ects of genetic regulation, of stage of growth and the interaction between these two factors. The correlations were also calculated with the GLM procedure. RESULTS Cell wall composition Compared to that of controls, the proportion of CWRs from the transgenic samples was similar in lines 40-1 and 50 but signiÐcantly lower in lines 40-2-8 and 48 (P \ 0É05) (Table 1). The proportion of CWRs in the stems also decreased from the base upwards and was lower (P \ 0É01) in the young (40-2-8) whole stem than in the older (40-2-14) plants. No di†erence was observed in gravimetric lignin contents (on CWR basis) between the controls and their transgenic counterparts. A signiÐcant decrease in ABL content was observed only between transgenic and control whole stems of samples 40-1. The lignin content of the stems was broadly similar from base upwards between control and transgenic older plants but decreased from about 154 g kg~1 CWR in the bases to 95 g kg~1 CWR in the tops of the young 40-2 stems (data not shown). When measured gravimetrically, the cellulose contents in transgenic samples were broadly similar to those in controls (except in samples 40-2-14) whereas non-cellulose polysaccharide contents showed irregular variations. When estimated as reducing sugars (in acid hydrolysates), the non-cellulose polysaccharide and cellulose contents in stems of lines 40-1 and 50 were lower than determined by the gravimetric method (on average by 90 and 68 g kg~1, respectively) and were found to be higher in the stems of the two transgenic lines than in the controls (on average 340 g kg~1 compared to 310 g kg~1). Uronic acid contents were unchanged between control and transgenic samples (on average 55 ^ 4, 59 ^ 9, 21 ^ 3 g kg~1 CWR for lines 40-1, 50 and 48, respectively). TABLE 1 Cell wall content and composition of the samples Samplea 40-2-8 control 40-2-8 AS 40-2-14 control 40-2-14 AS 40-1 control 40-1 AS 50 control 50 AS 48 control 48 AS RSE CW Rb (g kg~1 DM) WS WS WS WS WS T M B WS T M B WS T M B WS T M B WS WS 423 393* 611 605 767 656 668 746 767 682* 741* 747 723 638 711 788 726 659 704 771 767 712* 9É9 Cell wall constituents (g kg~1 CW R)c L ignin monomers (g kg~1 JL )d JL ABL HC C V 80 85 123 116 148 140 146 145 147 134 147 150 155 156 155 152 154 154 156 151 156 160 4É4 202 165 207 192 265 200 209 232 229* 217 206 233 241 210 215 207 229 223 212 221 161 186 15É9 457 452 427 452* 421 439 433 419 422 421* 418 401* 409 377 403 387 402 394* 418* 404* 371 370 7É4 457 462 447 425* 422 418 414 426 426 441 434 435 429 460 442 445 433 451 420 355 473 470 6É2 227 190* 210 170* 214 223 202 212 204 208 215 225 180 190 133 198 187 183 143 207 138 122 9É7 S 184 141* 168 121** 196 188 187 202 173* 175 178 190 157 168 123 176 149 149* 105* 165 67 45* 9É4 S/V e NBO 414 337** 380 293* 413 416 393 416 380 386 392 418 340 360 259 376 340 335 250 375 348 349 16É9 0É68 0É62* 0É67 0É59** 0É77 0É70 0É77 0É80 0É71* 0É70 0É69* 0É71* 0É73 0É74 0É77 0É74 0É67* 0É68* 0É61** 0É67* 0É41 0É31** 0É02 a AS, transgenic (antisense) sample ; WS, whole stem ; T, top internodes ; M, middle internodes ; B, basal internodes ; RSE, residual standard error. b CWR, cell wall residue ; DM, dry matter. c JL, Jarrige lignin ; ABL, acetyl bromide lignin ; HC, hemicelluloses (determined by gravimetry) ; C, cellulose (determined by gravimetry). d V, vanillin ; S, syringaldehyde ; NBO, nitrobenzene oxidation products \ V ] S ] p-hydroxybenzaldehyde. e Molar ratio. *, ** SigniÐcant di†erences between data of transgenic and control counterpart at P \ 0É05, P \ 0É01, respectively. Cell wall composition and degradability of tobacco stems 509 EsteriÐed phenolic acids were not detected in the CWRs. Except in line 40-1, the ASCWP fraction was greater in transgenic CWRs than in controls (241, 136, 182, 202, 265 g kg~1 CWR compared to 230, 115, 124, 175, 189 g kg~1 in lines 40-1, 40-2-8, 40-2-14, 48 and 50, respectively) (Bernard Vailhe et al 1996). The aldehydes, resulting from nitrobenzene oxidation of the uncondensed monomers from lignins, were mainly vanillin and syringaldehyde with a small amount of phydroxybenzaldehyde. The total aldehyde content amounted to between 290 and 410 g kg~1 of JL in the di†erent samples (Table 1). The proportion of syringaldehyde from lignins was generally lower in the transgenic samples than in the control fractions, the lowest changes being observed in the sections of stems from line 40-1. Except in line 40-2, where it decreased, the proportion of vanillin was unchanged. The S/V molar ratio decreased in all samples, with the greatest di†erence being observed in line 48. The NBO content decreased signiÐcantly only for samples from line 40-2, which indicates an increase in the “condensedÏ carbon to carbon inter-residue lignin linkages. Degradability The extent of ISDMD of control stems was generally low with little variation (36È39%) between the 16-weekold plants (40-1, 48 and 50) (Table 2), and was slightly greater than that of the standard wheat straw. It increased up to 73% (stems of 40-2-8) for younger plants. A gradient of increasing ISDMD was observed in the internodes of 40-1 and 50 from the base upwards. Compared to that of control line 40-1, the extent of ISDMD of transgenic whole stems was unchanged. However, some di†erences were observed between the corresponding fractions in the internodes, probably because of inhomogeneity in the internode sampling. Since the plants had a variable number and size of TABLE 2 In situ disappearance (48 h) and cellulase solubility of initial dry matter and cell wall residue (CWR) of the tobacco samples Samplea 40-2-8 control 40-2-8 AS 40-2-14 control 40-2-14 AS RSE 40-1 control 40-1 AS RSE 50 control 50 AS RSE 48 control 48 AS RSE In situ disappearance (%) WS WS WS WS WS T M B WS T M B WS T M B WS T M B WS WS Cellulase solubility (%) Dry matter CW R Dry matter CW R 73É0 77É3** 47É3 51É4* 1É16 38É7 50É5 46É2 26É8 37É9 45É7** 39É2** 28É8** 0É50 36É5 46É7 35É8 27É0 41É2** 47É0 38É2** 29É4** 0É52 38É5 45É3** 0É54 36É2 42É2** 13É8 19É7* 2É02 20É1 24É6 19É5 1É9 19É0 20É4** 17É9** 4É7** 0É73 12É1 16É4 9É6 7É3 19É0** 19É6** 12É3** 8É4* 0É73 19É8 23É2** 0É86 72É5 78É9** 48É4 51É4** 0É70 40É2 51É5 46É4 30É5 39É0 49É0 41É4** 34É4** 1É06 36É2 44É7 38É5 28É6 40É4** 49É2** 42É0** 33É7** 0É82 35É4 41É6** 0É30 35É0 46É3** 15É5 19É7** 1É53 22É0 26É1 19É7 6É8 20É5 25É2 20É9 12É2** 1É49 11É8 13É2 13É5 9É3 17É8** 22É8** 17É6** 14É0* 1É14 15É8 18É0* 0É40 a AS, transgenic sample ; WS, whole stem ; T, top internodes ; M, middle internodes ; B, basal internodes ; RSE, residual standard error. *, **, *** SigniÐcant di†erences between data of transgenic and control counterpart at P \ 0É1, P \ 0É05, P \ 0É01, respectively. M A Bernard V ailhe et al 510 internodes, it was not easy to make a comparable sampling of upper young tissues and middle and basal older tissues for all the plants. In contrast, the extent of ISDMD of transgenic batch 40-2 was signiÐcantly greater than the controls in both the 8-week and 14-week old plants, independent of the stage. Similarly, for the other transgenic lines, the extent of ISDMD was signiÐcantly greater than for the controls in all fractions studied except for the upper part of the stems of line 50. The greatest di†erences were observed in lines 50 and 48 (increase of 4É8 and 6É8 percentage units in whole stems). Except for samples 40-1, the CWR was degraded to a greatest extent in all fractions of the transgenic stems, with the highest increase (7 percentage units) observed in line 50. In transgenic line 48, the increase in ISCWD was lower than that of ISDMD, owing to the greater cell soluble content of the transgenic sample. The variations of cellulase digestibility were roughly similar to those of in situ disappearances (Table 2). Both methods were in good agreement : for 22 samples, CDMD \ 0É93]ISDMD ] 4É12, RSE \ 2É3, R \ 0É98). In contrast to the in situ Ðndings, even the tops of transgenic line 50 were more degradable than those of the control. In addition, the enhanced degradability due to genetic modiÐcation was signiÐcantly (P \ 0É05) lower in the older stems of 40-2 than in the younger stems. Non-cellulose polysaccharides were more degradable than cellulose (Table 3). Except for line 40-1, improvements in ISCWR degradability in transgenic stems were due to a greater degradation of all polysaccharides. Surprisingly, disappearances of Jarrige lignin in the rumen gave negative values but, in lines 50 and 40-2-8, the transgenic lignin content of fermented residues was signiÐcantly lower than in controls. In addition, the NBO content decreased from 34 to 27% JL after in situ incubation of transgenic line 50 but remained unchanged in control line 50 and in both transgenic and control line 40-1. This indicates a di†erence in the solubilisation of lignins and an increase in the degree of condensation of the transgenic lignin after rumen incubation : the latter was mainly due to preferential losses of syringyl units. The kinetics of in situ disappearance clearly show the respective e†ects of soluble cell material and wall degradation on the pattern of the degradation curves of dry matter and of CWR. Improvement of ISCWD was particularly great in line 50 (Fig 1). The values obtained for the kinetic parameters (Table 4) showed a good correspondence between calculated and experimental curves. The indigestible fraction of the transgenic stems was signiÐcantly lower in all samples except line 40-1. The inÑuence of the soluble fraction was mainly seen in line 48 since less than half of the potential ISDMD (a ] b) increase in transgenic stem was due to enhanced cell wall degradation (b). Although for ISCWD the slowly degradable fraction (b) was not enhanced for the transgenic counterpart of line 40-1, a clear increase was measured in line 40-2, which had a lower residual CAD activity, without signiÐcant interaction between the transgenic modiÐcation and stage. The greatest increases in “bÏ were observed for transgenic lines 50 and 40-2-14. The rates of degradation (c) increased for transgenic lines 50 and 40-2-8 but not for 40-1, 40-2-14 and 48. The lag phases calculated from ISCWD Ðts were e†ectively zero or similar (about 1É6% h~1 in line 40-2-8) in control and transgenic samples, except in line 50, for which it was lower for the transgenic than for the control (0É9 and 3É2% h~1 respectively, P \ 0É01). Relationships between lignin modiÐcations and degradability The stems of transgenic line 48 had the greatest decrease in S/V and the greatest increase in ISDMD, TABLE 3 In situ disappearance (48 h) of the cell wall constituents in whole stems (% initial content) Samplea Hemicelluloses Cellulose L igninsb 40-2-8 control 40-2-8 AS 40-2-14 control 40-2-14 AS 40-1 control 40-1 AS 50 control 50 AS 47É2 49É6 25É3 30É0 29É3 32É9 19É5 31É2* 38É7 49É0 14É4 18É8 13É5 15É9 4É8 24É8* [45É1 [40É5 [30É4 [36É7* [3É4 [12É3 [10É4 [4É7 a AS, transgenic sample. b Jarrige lignins. *, ** SigniÐcant di†erences between data of transgenic and control counterpart at P \ 0É05 and P \ 0É01, respectively. Fig 1. In situ disappearances of dry matter (squares) and of cell wall residue (circles) in control (open symbols) and transgenic (closed symbols) whole stems of line 50. Each bar represents ^ 1 SD of the mean. Cell wall composition and degradability of tobacco stems 511 TABLE 4 Parameters of the kinetics of dry matter and CWR disappearances in whole stems Samplea 40-2-8 control 40-2-8 AS 40-2-14 control 40-2-14 AS RSE 40-1 control 40-1 AS RSE 50 control 50 AS RSE 48 control 48 AS RSE DM disappearanceb CW R disappearanceb a b c Ind 43É8 44É9*** 25É9 25É5 0É20 22É4 22É4 0É41 18É8 19É3 0É40 23É3 29É1** 0É79 30É0 32É2*** 21É7 25É0** 0É48 16É6 15É4 0É85 17É1 20É9** 1É08 14É6 16É9 1É36 0É15 0É17** 0É33 0É30 0É08 0É22 0É17 0É04 0É15 0É20 0É03 0É10 0É06 0É02 26É2 22É8*** 52É5 49É5** 0É44 61É0 62É2 0É55 64É1 59É7** 0É78 62É1 54É0*** 0É88 b 39É7 43É0* 14É9 21É7** 1É61 20É6 19É0 0É74 14É1 19É2** 1É07 19É0 23É8** 1É40 c 0É08 0É12 0É13 0É07 0É03 0É20 0É16 0É03 0É05 0É11** 0É01 0É10 0É07 0É02 a AS, transgenic sample ; RSE, residual standard error. b a, rapidly degradable material ; b, slowly degradable material ; c, rate of degradation (%, h~1) ; Ind, indigestible material. *, **, *** SigniÐcant di†erences between data of transgenic and control counterpart at P \ 0É1, P \ 0É05, P \ 0É01, respectively. but a relatively low increase in “bÏ. Line 50 had relatively lower changes in composition or lignin structure but a great increase in “bÏ. Line 40-2-14 was the only material which showed signiÐcant changes in cell wall composition and lignin structure as well as moderate increases in ISCWD. Indeed, poor correlations were observed for pooled samples between V, S or S/V and the parameters from ISDMD or ISCWD kinetics (“bÏ or “a ] bÏ) ; the best observed (R \ 0É78) was between the decrease in S/V and the increase in “a ] bÏ for whole stems, but it was not signiÐcant. However, for the mature plants in which CAD activity was the most altered (ie excluding 40-1 samples), a signiÐcant negative correlation was observed between S or S/V and ISCWD (R \ [0É59 and R \ [0É68, respectively, P \ 0É05, n \ 12). In addition, although only four paired samples were studied, for the population of mature stems, the increases in ASCWP paralleled the improvements in ISCWD (Fig 2). DISCUSSION Cell wall content and composition Fig 2. Variation across mature plants in (=) improvements of in situ cell wall disappearance (ISCWD) and (4) increases in yield of alkali-soluble cell wall polymers (ASCWP) due to the transgenic modiÐcation. Our results on lignin content and composition in transgenic plants with reduced CAD activity are consistent with those of a previous study (Halpin et al 1994), in which only lines 40-1 and 50 were studied. Reduction of CAD activity was expected to decrease the formation of hydroxy-cinnamyl alcohols and therefore reduce the lignin content. Lignin contents were not changed, but decreases in the S/V molar ratio and in NBO suggest changes in lignin composition and structure. The syringyl content of the lignin was reduced but guaiacyl content was a†ected to a lesser extent. This suggests a mechanism exists that di†erentially regulates the deposition of syringyl and guaiacyl units in lignin (Terashima and Fukushima 1989). Alternatively, enzymes, distinct from the CAD down-regulated here, may be able to catalyse the production of coniferyl 512 alcohol, as suggested by Boudet and Grima Pettenati (1996) for Eucalyptus sp. However, several factors support the probability that lignin in the transgenic plants incorporates compositional changes in addition to those identiÐed here. An increase in the alkaliextractibility of the transgenic lignins was found (Halpin et al 1994 ; Bernard Vailhe et al 1996). Greater amounts of vanillin and syringaldehyde could be extracted by mild alkali from cell wall samples of transgenic plants than could be extracted from control plants (Bernard Vailhe et al 1996 ; Chabbert and Monties, pers comm). Halpin et al (1994) reported that transgenic stem samples showed a deeper staining with phloroglucinol (reactive with cinnamaldehydes), compared with control samples, and larger amounts of coniferyl and sinapyl aldehydes when analysed by pyrolysis mass spectrometry. It has therefore been suggested that a by-pass of CAD occurred and that cinnamaldehydes were incorporated into lignins. Since their presence could not be detected by thioacidolysis (Halpin et al 1994), FTIR or 13C NMR (Bernard Vailhe et al 1996), it is possible that they are located in condensed moieties, such as pinoresinols, and thereby produce an altered lignin (Halpin et al 1994). In contrast to the data on bm1 maize (Muller et al 1971) and bmr6 sorghum (Cherney et al 1986), lignin content in our transgenic samples did not decrease, but other similarities with these plants were observed. A lowering of the CWR (Muller et al 1971), of the S/V molar ratio and an increase in the degree of condensation have also been noted in bm1 maize (Kuc and Nelson 1964 ; Kuc et al 1968 ; Gordon and Griffith 1973) and similar variations in lignin composition were observed in bmr6 sorghum (Chabbert et al 1993). It has already been suggested that cinnamaldehyde incorporation alters lignins in bmr6 mutant sorghum (Pillonel et al 1991), and perhaps in bm1 maize, which also had a brownish xylem (Gee et al 1968). Degradability No improvement in degradability was observed in transgenic population 40-1, which had the highest residual CAD activity. Population 50, with the lowest CAD activity, and 40-2-14, which was only slightly more depressed than 40-1, had the greatest improvement in CWR degradability. It is likely, therefore, that the variations were due both to the extent of CAD down-regulation and to the heterogeneity of the population (Halpin et al 1994). A threshold of over 80% reduction in CAD activity is probably necessary to achieve signiÐcant improvements in digestibility. Wide variations have also been reported for whole bm1 maize plants ; the increases in IVDMD were either insigniÐcant (Lechtenberg et al 1972 ; Gordon and Neudoer†er) or similar to those observed in this study (Barnes et al M A Bernard V ailhe et al 1971). The cellulase solubilities measured for the tobacco samples were in good agreement with the extent of in situ disappearances, and the improvements in susceptibility to cellulase in severely CAD downregulated transgenic samples were consistent with those observed for bmr6 sorghum (Pillonel et al 1991). A comparison between dicots and grasses should be made with caution, since there are structural di†erences between the two, such as the presence of ferulic acid bridges between lignins and polysaccharides in grasses. The main e†ect of genetic modiÐcation was to enhance CWR degradability but the impact on DM degradability depended on the concomitant variation in CWR content, which decreased in line 48. Consistent with this are the changes previously reported in the bmr6 sorghum mutant (Porter et al 1978). Negative lignin disappearances have been observed in other digestibility trials (Besle et al 1995). Greater lignin disappearances and higher syringyl solubilisation observed in some tobacco transgenic populations have also been reported for bm1 maize (Gordon and Neudoer†er 1973). Wide variations were also observed in the rate of degradation. The slight increase in rate observed for two lines and the absence of any signiÐcant di†erence in the other line are in agreement with similar variations in rates observed for bm1 maize (Cymbaluk et al 1973 ; Gordon and Neudoer†er 1973). A lower lag phase of the transgenic stems in line 50 than in the control, together with a higher rate of degradability, is predictive of a higher intake of the down-regulated plants. A slight increase in intake of bm1 maize was found in experiments on young bulls (Barrière et al 1994). The interaction between transgenic modiÐcation and stage of growth (line 40-2) was only signiÐcant for cellulase digestibility. The absence of e†ect of downregulation in population 40-1 was probably due to a lower decrease in CAD activity and to the variability between plants rather than to greater maturity (2 weeks older) than batch 40-2-14. Whatever the case, any e†ect of stage of growth needs to be conÐrmed. In bm1 and normal maizes, improvements in degradability due to mutation slightly decreased with maturity (Barnes et al 1971). Relationships between cell wall modiÐcations and degradability In this study, the improvements in degradability were related to S or S/V. For a similar lignin content, the decrease in the proportion of syringyl unit (or S/V) in the secondary wall may increase the accessibility of the glycanases to the substrate and therefore improve digestibility (Reeves 1987 ; Jung and Deetz 1993). If this did occur in CAD down-regulated plants, it would only constitute part of the explanation since the correlation Cell wall composition and degradability of tobacco stems 513 between syringyl units or S/V and degradability were only signiÐcant for a restricted population. Otherwise, the changes in proportion of ASCWP may derive from structural changes related to cinnamaldehyde incorporation, similar to those occurring in the bm1 maize mutation (Gordon and Neudoer†er 1973). Covalent intermolecular bonds may be a†ected, particularly those between lignin and carbohydrates, and hence degradability. Further studies are needed to understand which structures in addition to syringyl units in transgenic lignins have a causal e†ect on improvements in degradability. Except in rare cases (Sommerfeldt et al 1979), it is not possible to di†erentiate the e†ects of lignin content from those of lignin structure on digestibility in natural mutants. The data from plants with a severely depleted CAD activity provide Ðrm additional evidence of the e†ect of lignin structure on the extent, and possibly the rate, of cell wall degradability. Tobacco is not used in ruminant nutrition so the tobacco model plant has a limited relevance to feeding applications. It is not possible with such a model to predict the e†ects of genetic manipulations on forage plants. Nevertheless, genetic engineering may be a means of obtaining forages with improved digestibility but without reduction of lignin content. Besle J M, Lassalas B, Thivend P 1981 Digestion des glucides cytoplasmiques de la feverole par le veau preruminant. Reprod Nutr De velop 21 629È649. Besle J M, Cornu A, Jouany J P 1994 Roles of structural phenylpropanoids in forage cell wall digestion. J Sci Food Agric 64 171È190. Besle J M, Cornu A, Jouany J P 1995 Transformations of structural phenylpropanoids during cell wall digestion. FEMS Microbiol Rev 16 33È52. Blumenkrantz N, Asboe-Hansen G 1973 New method for quantitative determination of uronic acids. Anal Biochem 54 484È489. Boudet A M, Grima-Pettenati J 1996 Lignin genetic engineering. Molec Breeding 2 12È39. Bucholtz D L, Cantrel R P, Axtell J D, Lechtenberg V L 1980 Lignin biochemistry of normal and brown midrib mutant sorghum. J Agric Food Chem 28 1239È1241. Chabbert B, Tollier M T, Monties B 1993 Lignin variability among di†erent brown midrib sorghum lines. In : 7th Intern Symp W ood Pulp Chem. Vol 1, CTAPI, Beijing, China, pp 462È468. Cherney J H, Moore K J, Volenec J J, Axtell J D 1986 Rate and extent of digestion of cell wall components of brownmidrib sorghum species. Crop Sci 26 1055È1059. Cymbaluk N F, Gordon A J, Neudoer†er T S 1973 The e†ect of the chemical composition of maize plant lignin on the digestibility of maize stalk in the rumen of cattle. Br J Nutr 29 1È12. Cordesse R 1982 Amelioration de la valeur nutritive des pailles par les traitements chimiques. Thèse Doctorale, ENSA, Montpellier, France. Demarquilly C, Chenost M 1969 Etude de la digestion des fourrages dans le rumen par la methode des sachets de nylon, liaisons avec la valeur alimentaire. Ann Zootech 18 419È436. Dhanoa M S 1988 On the analysis of dacron bag data for low degradability feeds. Grass Forage Sci 43 441È444. Englyst H N, Wiggins H S, Cummings J H 1982 Determination of the non-starch polysaccharides in plant foods by gasÈliquid chromatography of constituent sugars as alditol acetates. Analyst 107 307È318. Gee M S, Nelson O E, Kuc J 1968 Abnormal lignins produced by the brown-midrib mutants of maize II. Comparative studies on normal and brown-midrib-1 dimethylformamide lignins. Arch Biochem Biophys 123 403È408. Gordon A J, Griffith P R 1973 Chemical and in vivo evaluation of a brown midrib mutant of Zea mays II. Nuclear magnetic resonance spectra of digested and undigested alkali lignins and undigested dimethylformamide lignins. J Sci Food Agric 24 579È587. Gordon A J, Neudoer†er T S 1973 Chemical and in vivo evaluation of a brown midrib mutant of Zea mays. I. Fibre, lignin and amino acid composition and digestibility for sheep. J Sci Food Agric 24 565È577. Grand C, Sarni F, Boudet A M 1985 Inhibition of cinnamylalcohol-dehydrogenase activity and lignin synthesis in poplar (Populus ] euramericana Dode) tissues by two organic compounds. Planta 163 232È237. Halpin C, Knight M E, Foxon G A, Campbell M M, Boudet A M, Boon J J, Chabbert B, Tollier M T, Schuch W 1994 Manipulation of lignin quality by downregulation of cinnamyl alcohol dehydrogenase. Plant J 6 339È350. Halpin C, Foxon G A, Fentem P A 1995 Transgenic plants with improved energy characteristics. In : Biotechnology in Animal Feeds and Animal Feedings, eds Chesson A & Wallace R J. VCH, Weinheim, pp 279È293. Hanna W W, Monson W G, Gaines T P 1981 IVDMD, total sugars and lignin measurements on normal and brown- ACKNOWLEDGEMENTS The authors thank Michel Fabre for his contribution in carrying out the in situ disappearance measures. This work was supported by the European Union under the ECLAIR programme AGRE 0021 OPLIGE (Optimisation of lignin in crop and industrial plants through genetic engineering). REFERENCES Barnes R F, Muller L F, Bauman L F, Colenbrander V F 1971 In vitro dry matter disappearance of brown midrib mutants of maize (Zea mays L). J Anim Sci 33 881È884. Barrière Y, Argillier O 1993 Brown midrib genes of maize : a review. Agronomie 13 865È876. Barrière Y, Argillier O, Chabbert B, Tollier M T, Monties B 1994 Breeding silage maize with brown-midrib genes. Feeding value and biochemical characteristics. Agronomie 14 15È25. Baucher M, Chabbert B, Pilate G, Van Doorselaere J, Tollier M T, Petit-Conil M, Cornu D, Monties B, Van Montagu M, Inze D, Jouanin L, Boerjan W 1996 Red xylem and higher lignin extractability by down regulating a cinnamyl alcohol dehydrogenase in poplar, Plant Physiol 112 1479È 1490. Bernard Vailhe M A, Cornu A, Robert D, Maillot M P, Besle J M 1996 Cell wall degradability of transgenic tobacco stems in relation to their chemical extraction and lignin quality. J Agric Food Chem 44 1164È1169. 514 midrib (bmr) sorghums at various stages of development. Agronomy J 73 1050È1052. Iiyama K, Wallis A F A 1990 Determination of lignin in herbaceous plants by an improved acetyl bromide procedure. J Sci Food Agric 51 145È161. Jarrige R 1961 Analyse des constituants glucidiques des plantes fourragères : I. Fractionnement des constituants de la membrane par les hydrolyses acides. Ann Biol Anim Biochem Biophys 1 163È212. Jarrige R 1980 Chemical methods for predicting the energy and protein value of forages. Ann Zootech 29HS 299È323. Jorgenson L R 1931 Brown midrib in maize and its linkage relations. J Am Soc Agron 23 549È557. Jung H J G, Deetz D A 1993 Cell wall ligniÐcation and degradability. In : Forage Cell W all Structure and Digestibility, eds Jung H G, Buxton D R, HatÐeld R D & Ralph J. Am Soc Agron, Madison, USA, pp 315È346. Knight M E, Halpin C, Schuch W 1992 IdentiÐcation and characterisation of cDNA clones encoding cinnamyl alcohol dehydrogenase from tobacco. Plant Molec Biol 19 793È801. Kuc J, Nelson O E 1964 The abnormal lignins produced by the brown-midrib mutants of maize. Arch Biochem Biophys 105 103È113. Kuc J, Nelson O E, Flanagan P 1968 Degradation of abnormal lignins in the brown-midrib mutants and double mutants of maize. Phytochemistry 7 1435È1436. Lechtenberg V L, Muller L D, Bauman L F, Rhykerd C L, Barnes R F 1972 Laboratory and in vitro evaluation of inbred and F2 populations of brown midrib mutants of Zea mays L. Agron J 64 657È660. Michalet-Doreau B 1990 InÑuence de la nature de lÏaliment sur lÏimportance des pertes en particules dans la mesure de la degradabilite in sacco de lÏazote des aliments. Reprod Nutr Develop 151SÈ152S. Moerschbacher B M, Noll U, Gorrichon L, Reisener H J 1990 SpeciÐc inhibition of ligniÐcation breaks hypersensitive resistance of wheat to stem rust. Plant Physiol 93 465È470. Monties B 1986 Recent advances on lignin inhomogeneity. In : T he Biochemistry of Plant Phenolics (Vol 25), eds Van Sumere C F & Lea P F. Oxford University Press, Oxford, pp 161È181. M A Bernard V ailhe et al Mosoni P, Besle J M, Maillot M P, Gorrichon L 1994 14C Labelling of lignins of normal and bm3 maizes for fermentation studies. J Sci Food Agric 64 145È154. Muller L D, Barnes R F, Bauman L F, Colenbrander V F 1971 Variations in lignin and other structural components of brown midrib mutants of maize. Crop Sci 11 413È415. ^rskov E R, McDonald I 1979 The estimation of degradability in the rumen from incubation measurements weighted according to rate of passage. J Agric Sci Camb 92 499È503. Pillonel C, Mulder M M, Boon J J, Forster B, Binder B A 1991 Involvement of cinnamyl-alcohol dehydrogenase in the control of lignin formation in Sorghum bicolor L. Moench. Planta 185 538È544. Porter K S, Axtell J D, Lechtenberg V L, Colenbrander V F 1978 Phenotype, Ðber composition and in vitro dry matter disappearance of chemically induced brown midrib (bmr) mutants of sorghum. Crop Sci 18 205È208. Reeves J B 1987 Lignin and Ðber compositional changes in forages over a growing season and their e†ects on in vitro digestibility. J Dairy Sci 70 1583È1594. Rexen B 1977 Enzyme solubilityÈa method for evaluating the digestibility of alkali-treated straw. Anim Feed Sci T echnol 2 205È218. SAS 1989 SAS UserÏs Guide : Statistics (version 6, 1st edn). SAS Institute Inc, Cary, NC, USA. Scalbert A, Monties B, Lallemand J Y, Guittet E, Rolando C 1985 Ether linkage between phenolic acids and lignin fractions from wheat straw. Phytochemistry 24 1359È1362. Sommerfeldt J L, Schingoethe D J, Muller L D 1979 Brownmidrib corn silage for lactating dairy cows. J Dairy Sci 62 1611È1618. Terashima N, Fukushima K 1989 Biogenesis and structure of macromolecular lignin in the cell wall of tree xylem as studied in microautoradiography. In : Plant Cell W all Polymers, Biogenesis and Biodegradation (ACS Symp Ser 399) eds Lewis N G & Paice M G. American Chemical Society, Washington, DC, USA, pp 160È168. Venverloo C J 1971 The lignin of Populus nigra L-cv “ItalicaÏ and some other Salicaceae. Holzforschung 25 18È24.