measurement of low molecular weight tannins: indicators of methanogenic toxic tannins

JOURNAL OF FERMENTATION AND BIOENGINEERING Vol. 69, No. 3, 148-153. 1990

Measurement of Low Molecular Weight Tannins" Indicators of Methanogenic Toxic Tannins

JAMES FIELD, I* G A T Z E L E T T I N G A , I AND LEO H. A. HABETS 2

Department o f Water Pollution Control, Wageningen Agricultural University, Bomenweg 2, 6703 HD, Wageningen, ~ and Paques Incorperated, P.O. Box 52, 8560 A B Balk, 2 The Netherlands

Received 14 November 1988/Accepted 16 January 1990

The effectiveness of several low molecular weight (MW) tannin measurement methods for indicating the tannin toxicity to methan bacteria was evaluated. The methanogenic toxicity of the low and high MW tannins from autoxidized bark extracts was studied by selective removal of MW fractions from the extract with active carbon adsorption and calcium precipitation treatments. The toxicity of the low MW tannin fraction and the nontoxicity of the high MW tannin fraction were demonstrated. The low MW tannin concentration, measured by HPLC and a method based on the loss of tannins by treatment with granular active carbon (AC), had a very close relationship with the methanogenic toxicity, but a poor relationship was found with the total tannin concentration. The low MW tannins detected by the HPLC and AC methods had similar peak area positions in HPLC chromatograms as the tannins that were adsorbed by polyamide (trisacryl GF05) gel beads. These gel beads have an exclusion limit of 3000 g - m o l - I indicating that this is the approximate MW boundary between toxic and non-toxic tannins.

Coniferous bark extracts are useful for the study of tannin toxicity. The methanogenic toxicity of these extracts is almost entirely due to the tannin fraction (1). The tannins of bark are predominant ly oligomers (2-5). According to the trends outl ined by the tannin theory, only the oligomeric forms of tannins are effective inhibitors of methane bacteria (6). The detoxification of coniferous bark extracts by their autoxidat ive polymerizat ion is due to the conversion of the oligomeric tannins to high molecular weight (MW) tannins (7). Therefore, the total tannin concentrat ion cannot be used to estimate the residual toxicity of autoxidized extracts, since only the oligomeric fraction is responsible for the toxicity.

The purpose of this study was to study the applicabil i ty of measuring the low M W tannin fraction as an indicator of the methanogenic toxicity. Methods were investigated to distinguish low and high M W fraction of the tannins. The abil i ty to distinguish between monomeric and oligomeric tannins of the low MW fraction is not con- sidered necessary since the bark tannins contain few monomer ic tannins. A method for measuring the low MW condensed tannins with a H P L C technique has previously been described (6). In this study, alternative methods were also investigated using adsorbants that were selective for the low MW compounds . The use of such methods could be advantageous by eliminating the need for a H P L C to measure the low MW tannins.

M A T E R I A L S A N D M E T H O D S

Materials Extracts were prepared from the bark of Norway spruce (Picea abies), using 18 g . / ~ air dried milled bark in 60°C water and shaking for 3 h. To avoid premature autoxidat ion of the extracts, 250mg- I 1 of ascorbic acid was added and the extraction was done with

* Corresponding author. Present address: Department of Chemical Engineering, Autonomous University of Barcelona, 08193 Bellaterra, Barcelona, Spain.

N 2 gas in the head space. Analytical and bioassays The analytical methods for

the measurement of COD, UV (215nm) and color (440 rim) are described in previous articles (1, 7, 8). The total tannin measurement was based on the adsorpt ion of soluble bark extract matter on an insoluble polyamide polymer called polyvinylpyrrol idone (PVP). The PVP method of tannin measurement is described in detail by Field et al. (1). In this study, 14.3 g P V P . I ~ with bark extract was shaken for l h at 30°C and filtered. The difference in COD, UV at 215 rim, and color at 440nm before and after adsorpt ion on PVP is taken as the total tannin concentrat ion expressed either as a COD concentration or as absorbance units of UV or visible light.

The method used for measuring methanogenic toxicity is described in previous articles (1, 7, 8). In this study, the assay mixtures were not shaken. The volatile fatty acid (VFA) substrate was supplied at 4 .2g C O D . I ~. The granular methanogenic sludge concentrat ion, represented as volatile suspended solids (VSS), was 1.4 g VSS. / ~, and 2 g . / ~ NaHCO3 was used as a buffer. A second VFA feeding was done after 14 d of exposure to the extract-containing media to measure the residual sludge activity in the replaced media with extract not present. The control (fed VFA only ) methanogenic activities were 764 and 952 mg C O D . g 1 VSS. d ~ during the second VFA feeding of experiments labelled part "A" and "B", respectively. All bioassay results reported in this study are the averaged data of duplicate run experiments.

Measuring the low molecular weight tannins Three methods were used to measure the low MW tannins distinctly from the total tannin concentrat ion: (i) the H P L C method, estimating the low MW tannin fraction from HPLC chromatograms; (it) the AC method, measuring the tannins that are adsorbed by granular active carbon; and (iii) the GG method, measuring the tannins adsorbed by a polyamide gel bead adsorbant , trisacryl GF05. In all cases, the high MW tannins are calculated from the difference between total tannins (measured by PVP) and

148

VoL. 69, 1990 MEASUREMENT OF LOW MOLECULAR WEIGHT TANNINS 149

the low MW tannins. H P L C m e t h o d The HPLC method has been previously

described (6). In this study, the peak area was detected at 280 nm. The low MW tannins are represented by the tannin peak area of less than 28 min retention time. The ratio of the low MW tannin peak area to total tannin peak area is multiplied by the total tannin concentrat ion (measured by PVP) based on COD, UV at 215 rim, color (440 rim) to obtain the low MW tannin concentration estimate.

A C m e t h o d The AC method was done by measuring the PVP-measureable tannins in both a membrane-filtered sample and a membrane-filtered exhaustively AC treated (28 g . l ~ granular AC for 2 h) sample. Details of the exhaustive AC treatment were described previously (1). The difference in the tannin contents measured by PVP of the untreated and AC treated samples is equal to the low MW tannins. This method assumes that granular active carbon is only able to adsorb the low MW tannins. The ability of active carbon to adsorb the predominantly oligomeric bark tannins was established (1). Additionally, a hydrolyzable tannin oligomer (gallotannic acid) and a condensed tannin monomer (catechin) were tested and adsorbed for 96.2 and 98.59/00, respectively when treated with AC (as solution of 1 g-I 1). The inability of active carbon to adsorb high MW tannins was indicated from the poor removal of high MW tannins by AC treatment of autoxidized spruce bark extracts (9).

G G m e t h o d Trisacryl GF05 gel beads obtained from Soci6t6 Chimique, Villeneuve-La-Garenne, France, were used as an adsorbant for low MW tannins. The gel beads are manufactured for gel permeation chromato- graphy and have an exclusion limit of 3000g.mol 1. Trisacryl GF05 is composed of an amide-containing polymer that offers sites for hydrogen bonding with the low MW tannins that penetrate the beads.

The bark extract sample was shaken together with 35.8 g . l ~ (dry weight) trisacryl GF05 beads for 1 h, the membrane-filtered filtrate was again treated with the same amount of trisacryl GF05 for another hour and then membrane filtered. A replicate of the sample was brought through the same procedure without any beads present. The difference in COD concentration (or UV or color) between the untreated and trisacryl GF05 bead treated samples is equal to the low MW tannins COD (or UV or color). Distilled water was also treated with the gel beads to correct for COD and UV contributions originating from the gel beads. The method is abbreviated "GG".

Treatment of bark extracts for testing toxicity A u t o x i d a t i o n The autoxidation was generally (unless

otherwise stated) done by aerating extracts at a high pH (11.5) for approximately one day as described in Field et al. (7).

P V P t r e a t m e n t Extracts were shaken with 14.3 g-l 1 PVP for 1 h and then paper filtered.

A C t r e a t m e n t Extracts were intensively shaken with 28 .6g . / 1 granular AC for 2 h then centrifuged followed by paper filtering.

Ca 2 ~ t r e a t m e n t Extracts were treated at pH 8 with 500rag. /-1 of Ca 2~ (as CaC12) for 2h, then decanted through a paper filter.

G G t r e a t m e n t Experiments were attempted by treating the bark extracts with trisacryl GF05 gel under same conditions as the GG method for measuring low MW tannins. However, the gel introduced toxicity in the liquid which came into contact with it, most likely related to the presence of sodium azide in the gel (present in the

supplied product). For this reason, the bioassay experiments were abandoned for investigating the removal of toxicity with trisacryl GF05 gel.

RESULTS

Effects of treatments on COD, UV, color, and total tannins The effects of various treatments on the basic parameters of the bark extracts are shown in Table 1. Autoxidat ion was responsible for drastically increasing the color of the extract, which resulted from the polymeriza- tion of the tannins. Only minor changes in the total extract COD and UV were observed. The autoxidation did cause a distinct decrease in the total tannin concentration (approximately 30%o based on tannin COD and UV). The AC treatment removed 87 and 92%o of the total extract COD and UV, respectively, as well as removing 919/oo of the total tannin COD and UV. The AC treatment was thus capable of almost completely adsorbing the unoxidized bark matter. However, once the bark was autoxidized, the AC treatment was only able to adsorb 59 to 54O/oo of the total COD and UV, and only 40%o of the total tannin COD and UV. The lower adsorption of autoxidized bark matter on AC is perhaps due to a decreasing effectiveness of granular AC towards compounds of increasing MW(10). In contrast to AC, precipitation with calcium (Ca) was ineffective in eliminating total extract COD and UV as well as total tannin COD and UV from the unoxidized extract (only 7 to 149/00 elimination). The Ca precipitation was effective with the autoxidized extract, in which case 29 and 5 1 ~ of the total COD and UV, respectively, were precipitated and 30 and 41%0 of the total tannin COD and UV, respectively, were precipitated. The improvement is due to an increasing effectiveness of precipitation reactions with phenolic compounds of increasing MW (10-12).

TABLE 1. The COD, UV (215 nm), and color (440 nm) of untreated and treated spruce bark extracts and the total

tannin concentrations based on COD, UV, and color measurements

Extract COD(mg./ ~) UV(I~ , lcm) Color(l.,lcm)

No. Treat. a Total Tannin Total Tannin Total Tannin

A. AC Exp. b 2 untrt 4237 2397 109.5 88.4 0.51 0.24 3 untrt +AC 543 205 8.4 7.6 0.10 ND ¢ 4 auto 4127 1808 94.3 60.1 4.74 3.67 5 auto + AC 1679 1072 43.4 35.8 2.92 2.67 7 auto + PVP 2319 0 34.2 0.0 1.07 0.00

B. Ca Exp. u 2 untrt 4533 2881 139.5 110.1 0.76 0.51 8 untrt+Ca 4120 2606 120.3 102.9 0.43 0.33 4 auto 4372 1877 124.6 75.7 6.20 4.03 9 auto+Ca 3099 1316 60.9 44.8 2.88 2.46 7 auto+ PVP 2495 0 48.9 0.0 2.17 0.00

Treatments: 2, untreated spruce bark extract; 3, extract treated with the exhaustive AC method; 4, autoxidized extract (16h, pH 11.5); 5, autoxidized extract treated with the exhaustive AC treatment; 7, autoxidized extract treated with PVP treatment; 8, extract treated with Ca treatment; 9, autoxidized extract treated with Ca treatment.

b The spruce bark extracts were diluted 1.6 ~ (A) or 1.5 × (B) to a stock extract for the analysis, the results given in this table are the concentrations measured in the diluted extract. The extracts used in the AC (A) and Ca (B) experiments were prepared separately.

Not determined.

150 FIELD ET AL. J. FERMENT. B~OENG.,

TABLE 2. The low molecular weight (MW) tannins in untreated and treated spruce bark extracts as measured by various methods

Extract COD (mg.l ') UV (215 nm 1 ×, I cm) Color (440 nm 1 ~, 1 cm)

No. Treat? HPLC ~ GG' AC ~ AVG d HPLC GG AC AVG HPLC GG AC AVG

A. AC Exp. u 2. untrt 1889 1665 2192 1915 69.7 63.7 80.8 71.4 0.19 0.29 ND ~ 0.24 4. auto 591 840 736 722 19.6 18.3 24.3 20.7 1.20 0.74 1.00 0.98 5. auto+AC 144 149 NA e 147 4.8 4.8 NA 4.8 0.36 0.44 NA 0.40

B. Ca Exp. u 2. untrt 2161 1844 2751 2252 82.6 80.7 102.9 88.7 0.38 0.40 0.32 0.37 4. auto 460 346 638 481 18.6 18.0 28.2 21.6 0.99 0.65 1.30 0.98 9. auto+Ca 360 459 818 546 12.3 12.6 27.2 17.4 0.67 0.73 1.34 0.91

Treatment name abbreviations are defined in the footnote of Table 1. b See footnote in Table 1.

Low MW tannins: HPLC, estimated from the HPLC data; GG, measured by adsorption on trisacryl GF05; AC, measured by PVP tannin measurements before and after AC treatment of extract.

d Averaged low MW tannin concentration based on HPLC, GG, and AC methods. Not applicable.

t Not determined.

Effects of treatment on low and high MW tannins Table 2 and 3 present the low and high M W tannins measured by the three different me thods in the un t rea ted and t reated bark extracts . The tables i l lustrate that the ma- jo r effect o f the au tox ida t ion is to decrease the low M W tann in concen t ra t ion . The average decrease was 70 and 74%0 of the low M W tann in C O D and UV. The decrease in low M W tannins co inc ided with an increase in the high M W tannin concen t r a t ion . There fo re , the po lymer i za t ion react ions o f the au tox ida t i on change the M W c o m p o s i t i o n o f the to ta l tannins .

The res idual low M W tannins in the au tox id ized extract were highly adsorbed by the A C t rea tment . The average low M W tann in r emova l by A C was 79, 75, and 55°/00 based on C O D , UV, and color , respect ively. In contras t , the high M W tannins were no t effectively adsorbed . The average high M W tannin r emova l was 14, 25, and 3%o based on C O D , UV, and color , respect ively. The Ca precipi ta- t ion was not at all effective in prec ip i ta t ing the low M W tannins f r o m the au tox id ized extracts . H o w e v e r , the high M W tannins were e l imina ted on the average for 46, 50, and 500/00 based on C O D , UV, and color , respect ively, by Ca prec ip i ta t ion . In conc lus ion , the results indicate that the A C t r ea tmen t o f au tox id ized extracts selectively r emoves the low M W tannins , but Ca prec ip i ta t ion selec- t ively r emoves high M W tannins . These t rends are seen in the H P L C c h r o m a t o g r a m s o f Figs. 1 and 2.

The three different me thods o f measur ing low M W tannins , basical ly predic ted the same trends in the low M W tann in concen t r a t i on with respect to the var ious t rea tments o f the bark extract . The me thods only differed to a small extent in the absolu te concen t r a t ion o f low M W tannins measured . Genera l ly , the A C m e t h o d predic ted the highest concen t ra t ion , but the G G m e t h o d predic ted the lowest concen t ra t ion . F igure 1 i l lustrates that the A C m e t h o d exhaust ive ly r emoved the low M W tannin peak area in the H P L C plots o f the au tox id ized extracts, but also i l lustrates that the A C m e t h o d addi t iona l ly r emoved a small par t o f the high M W tannin peak area. The G G m e t h o d did no t exhaus t ive ly r e m o v e the low M W tannin peak area o f the H P L C plots , but it did not adsorb any o f the high M W tann in peak area in the au tox id ized extract .

Effects of treatments on methanogenic toxicity Figure 3 i l lustrates the effects o f the var ious bark extract t reat- ments on the me thanogen i c act ivi ty o f s ludge exposed to these extracts for 2 weeks. The act ivi ty results are plot-

ted together with a C O D balance o f the bark extracts to i l lustrate which f rac t ion o f the C O D is most likely responsible for the toxicity. The non - t ann in f rac t ion o f the autoxidized extracts are not responsible for any toxicity in accord- ance with our prev ious study (1). The au tox ida t ion o f the extracts dramat ica l ly reduces the toxicity, but part ial inhibi-

2> r i

i A

~ : i I!

4 I'2 ' 20 ' 2'8 3'6 ' Retention ~ime (rain)

FIG. 1. The HPLC chromatograms of untreated extract (A) and autoxidized (B) extracts of spruce bark. The boundary between low and high MW tannins is at 28-min retention time. For each sample, the top solid line is the entire extract, the bottom solid line is the PVP treated extract. The tannin peak area is the shaded area between the two chromatograms. The upper dotted line (long dashes) is the GG treated extract. The lower dotted line (short dashes) is the AC treated extract.

VoL 69, 1990 MEASUREMENT OF LOW MOLECULAR WEIGHT TANNINS 151

TABLE 3. The high molecular weight (MW) tannins in untreated and treated spruce bark extracts as measured by various methods

Extract COD (mg./ ') UV (215 nm 1 ×, 1 cm) Color (440rim I •, 1 cm)

No. Treat? HPLC c GG c AC ~ AVG d HPLC GG AC AVG HPLC GG AC AVG

A. AC Exp. 2. untrt 508 732 205 482 18.7 24.7 7.6 17.0 0.05 0.04 ND ~ 0.01 4. auto 1217 968 1072 1086 40.5 41.8 35.8 39.4 2.47 2.94 2.67 2.69 5. auto +AC 928 923 NA e 926 31.0 31.0 NA 31.0 2.31 2.23 ND 2.27

B. Ca Exp. 2. untrt 720 1037 130 629 27.5 29.4 7.2 21.4 0.13 0.11 0.19 0.14 4. auto 1417 1531 1239 1396 57.1 57.7 47.5 54.1 3.04 3.38 2.73 3.05 9. auto+Ca 956 857 498 770 32.5 32.2 17.6 27.4 1.78 1.73 1.12 1.54

a Treatment name abbreviations are defined in the footnote of Table 1. b See footnote in Table 1. c High MW tannins [(total tannin) (low MW tannin)]: HPLC, estimated from the HPLC data; GG, measured by adsorption on trisacryl

GF05; AC, measured by PVP tannin measurements before and after AC treatment of extract. o Averaged high MW tannin concentration based on HPLC, GG, and AC methods. e Not applicable. f Not determined.

tion is still evident. The partial inhibit ion cannot be due to the high MW tannins as these increase in concentrat ion as a result of the autoxidation. Therefore, the most likely fraction associated with residual toxicity in autoxidized spruce bark extracts is the low MW tannin fraction. The Ca precipitation, which has no effect on the low MW tannin fraction in autoxidized extract also had no effect in reliev- ing the residual toxicity even though it was capable of remov-

o I J J I I I r p I

B

0 9

A ° ¢ . -

i!iiiiiii!i!iii!i!i -u- ~ ~iiiiiiiililiiiii~~ iiiiiiiiiiiiiii!iiiii~ "~

i!iiiiii!iii!i!iii!iii~ "5

L, 12 20 28 36 Retention time (min)

FIG. 2. The HPLC chromatograms of autoxidized (A) and Ca treated autoxidized (B) extracts of spruce bark. The boundary between low and high MW tannins is at 28-min retention time. For each sample, the top solid line chromatogram is the entire extracts, the bottom solid line chromatogram is the PVP treated extract. The tannin peak area is the shaded area between the two chromatograms.

ing about half of the high MW tannins. On the other hand, the AC treatment of the autoxidized extract removed a ma- jority of the low MW tannin fraction without altering the high MW tannin concentration and this corresponded to a distinct decrease in the methanogenic toxicity.

Tannins as an indicator of methanogenic toxicity The data on tannins measured by PVP from previous studies and this study are plotted in Figs. 4A and 4B as a concentrat ion in the toxicity assay media versus the activity of the exposed sludge. When the results were obtained from an experiment with a dilution series of unoxidized bark extract, a strong relationship between the sludge activity and tannins measured by PVP was observed (Fig. 4A). From this experiment, the 50%o inhibiting concentration of spruce bark tannins was estimated to be 550 mg COD- l - ~. In contrast, when the results obtained from experiments where the bark extracts were oxidatively treated or otherwise treated with AC or Ca 2~ , than the relationship between the activity and tannins measured by PVP was very poor (Fig. 4B).

The dilution experiments did not change the ratio of low MW to total tannins, while the oxidation experiments did. When the oxidation treatments alter the MW composition of the tannins, the total tannins are no longer good indicators of the methanogenic toxicity.

The HPLC-estirnated low MW tannin data are plotted in Fig. 5A as a concentrat ion in the toxicity assay media versus the activity of the exposed sludge. The results obtained from the oxidation experiments showed a strong relationship between activity and low MW tannins. This can be at- tributed to the fact that only the low MW tannins are truly toxic and in this case the high MW tannins are excluded from the detection. Of the 24 experiments, 2 points (en- circled in the graph) deviated to some extent from the average relationship found with the oxidized extracts. This was possibly due to the fact that these 2 points are from experiments that used extracts prepared from a different (older) sample of spruce bark.

The 50%o inhibiting concentration of the low MW tannins measured by HPLC from unoxidized extracts was 430mg COD./ -1 and from the oxidized extracts was 200 mg C O D . l - 1. The higher toxicity of low MW tannins from the oxidized extracts compared with those from the unoxidized extract indicates that the low MW tannins from these sources differ in toxicity. This behavior is predicted by the tannin theory which postulates an increas-

152 FIELD ET AL. J. FERMENT. BIOENG.,

1 0 0 - ~ ~ ] (AcC~Dp.)] 1 0 0 - - - COl) B a l a n c e B a l a n c e

--~ 80 80- (Ca exp.) .~

so , s0- YZ,

~ 40- 40-

o 20 - 20-

iiiiiil 0 0 - 2 4 5 7 2 4 9 7

120 120 Act ivi ty 777 Activit ,

100 (AC exp.) / / / 100 (Ca exp. ~ / /

o 80 ~ 80

~7o. 60 ~ /7~ 60 ~ / ~ / 1 / / / / /

"< 0 ' 0 ~/'/ 2 4 5 7 4 9 7

E x t r a c t t r e a t m e n t E x t r a c t t r e a t m e n t

. ~ High MW ' t a n n i n COD ~ ~ Act ivi ty

~ L ~ Low MW t a n n i n COD

FIG. 3. The methanogenic activity of sludge exposed for two weeks to unoxidized (2), autoxidized (4), autoxidized + AC (5), autoxidized ~- PVP (7), and autoxidized + Ca (9) treated bark extracts. The activity is expressed as a percentage of the activity obtained with a VFA fed control. The results are compared with the low and high MW tannin and non-tannin fractions expressed as a percent of the total COD in the unoxidized extract (~ total tO). The stock extract concentrations reported in Tables 1, 2 and 3 were diluted 1.25 X in the assay media for Experiment A (AC) and diluted 1.16 X for Experiment B (Ca).

ing effectiveness of tannins as their size increases, at least as long as they are capable of penetrating bacteria (6). Based on this postulate, we would expect the tannins with the largest size in the low MW tannin fraction to be the most toxic. The low MW tannins still present after polymeriza- tion treatments are more likely to have a greater size than the low MW tannins of the original unoxidized extract.

The low MW tannin data measured by AC are plotted in Fig. 5B as a concentrat ion in the toxicity assay media versus the activity of the exposed sludge. These results, obtained from the oxidation experiments, also showed a strong relationship between toxicity and low MW tannin concentration. The same two points which deviated in the figure with the HPLC data, also deviated to the same extent with the AC low M W ' t a n n i n data which indicates that the anomalous behavior is most likely due to the extract rather than the measurement methods.

The 50%0 inhibiting concentrat ion of the AC low MW tannins from the oxidized extracts was 250 mg COD . / ~. This was slightly less toxic than the low MW tannins estimated from the HPLC method because the AC method also adsorbs to a small extent some high MW tannins. However the preferential adsorption for low MW tannins was high enough that this method is equally as useful for indicating the toxicity as the HPLC low MW tannins.

©

e,.* 3,

i

4..s ;D

1 0 0 - a

no I

6 ()

,ll)

2O

() 0

12.OK[

1 0 0 J

80

60

10

0

FIG. 4.

(A) J

• A

x A i

• A * A [ , . - ~ A

:500 1000 1500 2000 :2500

i•l I B ) y.

K tI " i D (,

. 2 ( ( ( • (

( C

d----Z6o 1ooo 1:,o oo

P V I ~ l ; - t t l l l i l l c l ' ~ I H ' C l l l l ' t t l i o l I { I i 1 ~ ( • ( ) [ ) / / £ )

The methanogenic activity of sludge after two weeks of exposures to spruce bark extracts versus the assay concentration (during the exposure period) of the tannins measured by PVP. (A) Data from dilution of the unoxidized extracts. (B) Data from oxidation treatments of undiluted extract• Experiments: 1 (±) - Field et al. (1); 2 (, : ) - Field et al. (7); 3 ( • ) and 4 ( [] ) = Field e t a / . (9); 5 ( • ) - and 6 (A)=this study (Fig. 3). Treatments: A, extract only; B, extract autoxidized (< 15min); C, extract autoxidized (:.15min); D, extract treated 4 weeks biologically (aerobic); E, extract oxidized with H~O2; E', autoxidized extract oxidized with H202; F, extract treated half exhaustively with AC; G, autoxidized extract treated half exhaustively with AC; H, autoxidized extract treated exhaustively with AC; 1, autoxidized extract treated with Ca; J, extract treated with PVP; K, autoxidized extract treated with PVP.

DISCUSSION

The low MW fraction of tannins present in bark extracts was demonstrated to be the only tannin fraction responsible for toxicity to methanogenic bacteria. The high MW tannin fraction produced by the oxidative treatments of the extract was likewise shown to not inhibit methane bacteria. Therefore an effective indication of tannin toxicity in bark extracts should limit itself to the detection of only the low MW tannin fraction as opposed to the total tannins. The low MW tannin concentration, measured by the HPLC and AC methods, had a very close relationship with the methanogenic toxicity, but a poor relationship was found, based on the total (PVP) tannin concentration. The low MW tannins detected by the HPLC and AC method had the same peak area positions (in the HPLC chromatograms) as those that were adsorbed by polyamide gel beads. The adsorption sites of these beads are limited in availability to tannins with a MW less than 3000. This MW constitutes the approximate upper limit of the various low MW tannin determinations proposed and likewise serves as an approximate boundary between non-toxic and toxic tannins. This MW has also been suggested to be the maxi- mum effective tannin size for the tanning of hides (13).

The AC method of low MW tannin measurement is recommended most highly over the other methods. It would be the least expensive because chromatographic equipment

Vor. 69, 1990 MEASUREMENT OF LOW MOLECULAR WEIGHT TANNINS 153

©

b © O

1 20 .~JE '

100~,i~ E K

8O

6O

40

2O

0 0

120

(a)

~u

/

7

500 1000 15"~0 2~ 30

...a

<i

1oo 80

60

40

2O

0

(B)

\ \ c

. . . . ~ ~ 0 500 1000 1 5 0 0 " " - - 2000

LMW t a n n i n c o n c e n t r a t i o n ( r a g C O D / l )

FIG. 5. The methanogenic activity of sludge after two weeks of exposures to spruce bark extracts versus the assay concentration (during the exposure period) of the low MW tannins. CA) Based on low MW tannin data of the HPLC method. (B) Based on the low MW tannin data of the AC method. All the data in this figure were obtained from oxidation treatments of undiluted extract. Experiment and treatment legends are in caption of Fig. 4. Experiments 7 ( -- ) and 8 (*) are additional ones of this study.

would not be necessary. Additionally, the granular active carbon and insoluble polyvinylpyrrolidone adsorbants are widely available.

REFERENCES

1. Field, J. A., Leyendeckers, M.J., Sierra-Alvarez, R., Lettinga, G., and Habets, L, H. A.: The methanogenic toxicity of bark tannins and the anaerobic biodegradability of water soluble bark matter. Wat. Sci. Tech., 20(1), 219-240 (1988).

2. Porter, L. J.: Extractives of Pinus radiata bark. II. Procyanidin constituents. New Zealand J. Sci., 17, 213-218 (1974).

3. Karehesy, J . J . and Hemingway, R.W.: Loblolly pine bark polyflavonoids. J. Agric. Food Chem., 28, 222-228 (1980).

4. Hemingway, R. W., Yeap Foo, L., and Porter, L.J.: Polymeric proanthocyanins: interflavonoid linkage isomerism in (epicatechin-4)-(epicatechin-4)-catechin procyanidins. J. Chem. Soc. Chem. Commun., 320-322 (1981).

5. Hemingway, R.W., Yeap Foo, L., and Porter, L.J.: Linkage isomerism in trimeric and polymeric 2, 3-cis-procyanidins. J. Chem. Soc. Perkin I, 1209-1221 (1982).

6. Field, J. A., Kortekaas, S., and Lettinga, G.: The tannin theory of methanogenic toxicity. Biological Wastes, 29, 241-262 (1989).

7. Field, J. A., Lettinga, G., and Habets, L. H. A.: Oxidative detoxification of aqueous bark extracts. I. Autoxidation. J. Chem. Tech. Biotechnol., in press (1990).

8. Field J .A . and Lettinga, G.: The methanogenic toxicity and anaerobic degradability of a hydrolyzable tannin. Wat. Res., 21, 367-374 (1987).

9. Field, J.A., Sierra-Alvarez, R., Lettinga, G., and Habets, L. H.A.: Oxidative detoxification of aqueous bark extracts. I1. Alternative methods. J. Chem. Tech. Biotechnol., in press (1990).

10. EI-Rehaili, A. M. and Weber, Jr., W.J . : Correlation of humic substance trihalomethane formation potential and adsorption behavior to molecular weight distribution in raw and chemically treated waters. Wat. Res., 21, 573-582 (1987).

11. Schmidt, R.L. and Joyce, T.W.: An enzymatic pretreatment to enhance the lime precipitability of pulp mill effluents. Tappi, 63(12), 63-67 (1980).

12. Miistein, O., Haars, A., Majcherezyk, A., Trojanowski, J., Tautz, D., Zanker, H., and Huttermann, A.: Removal of chlorophenols and chlorolignins from bleaching effluent by com- bined chemical and biological treatment. Wat. Sci. Tech., 20 (1), 161-170 (1988).

13. White, T.: Tannins-their occurence and significance. J. Sci. Food. Agric., 8, 377-384 (1957).

measurement of low molecular weight tannins: indicators of methanogenic toxic tannins

Documents