Effect of Pesticide Inoculation, Duration of Composting, and Degradation Time on the Content of Compost Fatty Acids, Quantified Using Two Methods

Abstract

Compost biobeds can promote biodegradation of pesticides. The microbial community structure changes during the composting process, and simple methods can potentially be used to follow these changes. In this study the microbial identification (MIDI) and ester-linked (EL) procedures were used to determine the composition of fatty acid methyl esters (FAMEs) in composts aged 3 and 12 months, inoculated with 3 recalcitrant pesticides (azoxystrobin, chlorotoluron, and epoxyconazole and a coapplication of all three) after 0, 56, and 125 days of degradation. Pesticide persistence was high, and after 125 days the residue was 22 to 70% of the applied amount depending mostly on the composting age. Seventy-one FAMEs belonging to nine groups were detected. The EL method provided three times as many detections as did the MIDI method and was more sensitive for all FAME groups except alcohol. Thirty-six and five FAMEs were unique to the EL and MIDI methods, respectively. The extraction method was of importance. The EL method provided a higher number of detections for 57 FAMEs, and the MIDI method provided a higher number for 9 FAMEs, while the two methods were equal for 5 FAMEs; thus, the EL method provided a more uniform overall FAME profile. Effects of the other factors were not always clear. Inoculation with pesticide did not influence the FAME profile with the MIDI method, while it influenced cyclopropane and monounsaturated content with the EL method. Composting age and degradation time had an effect on some groups of FAMEs, and this effect was greater with the EL method. The use of some FAMEs as biomarkers to follow microbial community succession was likely influenced by the type of compost and other factors.

Plant protection has become a key factor in intensive agriculture, but the widespread use of pesticides can be toxic to nontarget organisms and can lead to ecosystem alterations. Formulated chlorotoluron (a herbicide), azoxystrobin (a fungicide), and epoxyconazole (a fungicide) are commonly used at the postemergence stage mainly for cereals, to control broad-leaved weeds and grasses and to control some foliar and soilborne diseases, respectively.

Most information on pesticide persistence is derived from registration documents, few studies appear in the scientific literature, and independent research is always suggested. A recent pesticide database (Footprint, 2007; http://www.eu-footprint.org/ppdb.html, accessed April 2010) reports half-life values (days) in soil, derived from laboratory studies at 20°C, of 73.5 for azoxystrobin, 59 for chlorotoluron, and 226 for epoxyconazole. However, a certain variability exists, and in other research half-lives are longer, i.e., more than 2 years (at 10°C) for epoxyconazole (4), 62 to 107 days for azoxystrobin (10), and 30 to 200 days for chlorotoluron (36), showing that significant degradation should be measured on a time scale of months or years (8, 9) according to the results of a recent review on pesticide persistence (3).

The use of such persistent pesticides on agricultural fields can have a high impact on the environment through leaching into aquifers and altering the microbial community and processes in the soil. Investigations of structural and functional changes in soil microbial communities have been done using several methods, including the fatty acid profiling method (14, 27, 29). This procedure has also been used as an early and sensitive method to investigate microorganism succession during the composting process (32), together with the composition of long-chain fatty acids (26, 28).

Recent studies show a short-term effect on fatty acids during composting, with the main decrease being between days 3 and 24 of composting (32), and also a fatty acid profile shift in months-long experiments (2).

Simple standardized methods to monitor changes in the microbial community structure may have the potential to follow microbial succession and determine compost maturity and quality. Methods to determine the fatty acid composition of the microorganisms have thus been developed, such as the microbial identification (MIDI) system and the ester-linked (EL) procedure (25, 29). The application of these methods to compost samples is recent, and little is known about the extent to which they can give comparable information (32).

Compost is an interesting medium for biodepuration systems such as biobeds (20) because the high content and availability of organic matter could have a significant effect on pesticide adsorption and dissipation (16, 34). The effectiveness of the degradative capacity of those systems can be studied using a worst-case scenario, i.e., using pesticides recalcitrant to degradation.

Some studies show that less mature compost is better able to absorb and degrade pesticides, due to greater carbon sources for microorganisms to cometabolize and remove pesticides (37). For example, in samples taken from composting facilities (33) it was found that a 6-month-old compost had a greater capacity to remove atrazine from solution, because of its higher organic carbon content, and this suggested that less mature compost may be better suited for environmental applications. However, since most compost is produced in big facilities, much of the compost can remain in the facility and mature, changing physical-chemical and microbiological properties (13) and consequently pesticide removal capacity.

The main goal of this laboratory study is to assess the evolution of microorganisms in compost aged 3 and 12 months after inoculation with three recalcitrant pesticides (azoxystrobin, chlorotoluron, and epoxyconazole, the persistence of which is also measured) and a coapplication of all three, through the changes in the fatty acid methyl ester (FAME) profile using two methods, the microbial identification (MIDI) and ester-linked (EL) methods.

MATERIALS AND METHODS

Compost substrate.

The tested substrate was a mixture of urban waste and garden compost that came from GESENU (compost production plant) in Pietramelina, Perugia, Italy. Composts aged 3 months and 12 months (designated 3 M and 12 M, respectively) were used for the experiment. Even if the microbial activity in the early stages of composting is of importance, only 3 M and 12 M composts were used because this is the age range of industrial composts delivered by production plants and allowed for agricultural use and biobeds. Physical and chemical characteristics of the substrates were as follows: for 3 M compost, pH (KCl), 7.8; density, 0.434; C, 30.2%; N, 2.4%; C/N ratio, 12.6; mineral matter, 42%; and organic matter, 58%; for 12 M compost, pH 8.4; density, 0.574; C, 29%; N, 3.1%; C/N ratio, 9.3; mineral matter, 46%; and organic matter, 54%. The 2 composts were sieved at 2 mm and brought to 60% of their water-holding capacity.

Pesticides.

The commercial products Amistar, Alpha Chlorotoluron 500, and Opus were used, containing 23.1% (wt/wt), 43.9% (wt/wt), and 5.4% (wt/wt) azoxystrobin [methyl (E)-2-{2-[6-(2-cyanophenoxy)pyrimidin-4-yloxy]phenyl}-3-methoxyacrylate], chlorotoluron [3-(3-chloro-4-methylphenyl)-1,1-dimethylurea], and epoxyconazole {(2RS,3SR)-1-[3-(2-chlorophenyl)-2,3-epoxy-2-(4-fluorophenyl)propyl]-1H-1,2,4-triazole}, respectively. Pesticide chemical characteristics are reported in Table 1.

TABLE 1.

Chemical properties and environmental parameters of the three studied pesticides (from Footprint, 2007 [http://www.eu-footprint.org/ppdb.html])

Chemical propertya	Unit	Azoxystrobin	Chlorotoluron	Epoxyconazole
Mol wt		403.4	212.7	329.8
Solvent solubility (acetone at 20°C)	g/liter	86	54	140
Water solubility	mg/liter	6.7	74	7.1
Melting point	°C	116	148.1	136.7
Boiling point	°C	360
Freundlich constant (Kf)		0.8-2.9	5.24	4.8-21.8
KOC	ml/g	207-594	108-384	280-2,647
DT50	Days	35.2-153.4	52-66	98-649

^aK_OC, organic carbon/water partitioning coefficient; DT₅₀, half-life.

Incubation study.

The experiment had a completely randomized 2 × 5 factorial design of two compost ages (3 M and 12 M) and five pesticide treatments (control, Amistar, Alpha Chlorotoluron 500, Opus, and a coapplication of the three pesticides; the treatments are here designated UNT, AZO, CHL, EPO, and ACE, respectively) with three replications. One kilogram of compost was amended with pesticides at the rate of 100 mg kg of active ingredient⁻¹ or as a control and incubated in the dark at 20°C. The pesticide rate was selected as being high enough to cause detectable differences in FAME content until the end of the sampling period. Moisture was maintained by the addition of the required amount of water at weekly intervals as determined by gravimetric analysis.

Analytical methods.

The treatments were sampled after 0, 4, 12, 31, 56, 89, and 125 days by taking three 20-g subsamples which were each analyzed for pesticides by adding 50 ml of methanol and shaking them for 1 h using a reciprocal shaker. The solutions were separated from the compost by centrifugation (3,490 × g for 15 min at 4°C) and analyzed by high-pressure liquid chromatography (HPLC). Percent recovery rates (means ± standard deviations) of the method were 87 ± 6.1 for AZO, 87 ± 8.1 for EPO, and 103 ± 5.8 for CHL. Analysis was performed using an Agilent 1100 series HPLC equipped with a C₁₈, 15-cm by 4.6-mm, 5-μm column. The operating conditions were as follows: solvents, water with 0.1% orthophosphoric acid and acetonitrile (66%/34%); flow rate, 1 ml/min; and run time, 35 min. Retention times were 7.7, 30.5, and 32.5 min for CHL, AZO, and EPO, respectively. The detection limit was 0.1 μg kg⁻¹ for each pesticide.

Fatty acid extraction by the EL FAME or MIDI method was done after 0, 56, and 125 days of incubation on 1 g of compost placed in a screw-cap test tube. The procedures used derived from that of a previous similar study (29) with slight modifications necessary for compost (32).

EL procedure.

(i) Mild alkaline methanolysis was performed by adding 15 ml of fresh 0.2 M KOH in methanol, and the samples were then mixed and incubated at 37°C for 1 h; (ii) neutralization was performed by adding 3 ml of 1 M acetic acid and by partitioning into an organic phase with addition of 10 ml of hexane followed by centrifugation; (iii) extraction steps were performed by transferring the hexane layer to a glass tube, evaporating it under a stream of N₂, and dissolving it in 0.5 ml of hexane-methyl tertiary butyl ether (MTBE) (1:1).

MIDI procedure.

(i) Saponification was performed by adding 5 ml of 3.75 M NaOH in MeOH-H₂O (1:1) and heating the samples in a 100°C water bath for 30 min; (ii) methylation was performed by adding 6 ml of HCl in aqueous methanol (6 M) and incubating the mixture in a water bath at 80°C for 10 min; (iii) extraction was performed by addition of 3 ml of hexane-MTBE (1:1) and by centrifugation for 10 min; (iv) sample cleanup was performed by washing with 3 ml of 0.3 M NaOH.

The organic phase extracted with both methods was transferred to gas chromatography (GC) vials for the GC analysis, performed using a Hewlett-Packard HP6890 series gas chromatograph fitted with an Ultra 2, 25-m × 0.2-mm, 0.33-μm column. The starting temperature was 170°C, rising to 310°C at the end of the sample analysis, which took approximately 20 min. The GC was controlled by the HP Chemstation software that enabled a comparison with a commercially available library and an in-house library. The GC used a commercially available calibration standard which was analyzed twice prior to each set of runs and at set intervals during the run sequence.

FAME nomenclature.

FAMEs were named in accordance with standard nomenclature and as in reference 32: the total number of carbon atoms, followed by a colon and the number of double bonds. The position of the first double bond is indicated by ω followed by the number of carbon atoms from the aliphatic end. The suffixes c and t refer to cis and trans isomers, respectively. Methyl branching at the iso and anteiso positions and that at the 10th carbon atom from the carboxyl end are designated by the prefixes i, a, and 10Me, respectively. The prefix or suffix cy denotes cyclopropane fatty acids. When present, the number of hydroxyl substitutions is also given. Groups of FAMEs were named in accordance with a previous similar study (25) with separations made between mono- and polyunsaturated fatty acids to allow the use of some individual FAMEs or FAME groups as biomarkers, as follows: Gram-positive bacteria, iso and anteiso branched fatty acids; Gram-negative bacteria, monounsaturated fatty acids; fungi, 18:2ω6,9c; actinomycetes, 10Me18:0; mycorrhizae, 16:1ω5c; protozoa, 20:4ω6,9,12,15c (5, 17, 22).

Statistical analysis.

Pesticide half-life was calculated according to a first-order kinetic degradation with the module nonlinear estimation of Statistica 7.1 (Statsoft Inc., Tulsa, OK).

The detection result for each FAME was expressed as a percentage of the total amount of FAMEs. Percentages were transformed by the arcsine square root transformation to reduce the nonnormality of the data set distribution (6, 12), a reduction confirmed by distribution analysis. First, an analysis of variance (ANOVA) was performed for the factor extraction method. A main-effects ANOVA was then performed separately for each method in order to evaluate the effects of pesticide inoculation, composting age, and incubation time (here degradation time). Interactions were not considered difficult to interpret. All ANOVAs were performed for the individual and grouped FAMEs in order to obtain both an analytical and a synthetic assessment of the effects.

For each extraction method the correlations between the total contents of the various groups of FAMEs (the variables) and clustering of cases due to factors were analyzed using principal component and classification analysis to obtain a 2-dimensional representation of all information included in the data set, rescaling the factor coordinates of the cases with the highest distance to the axis origin.

Analyses were performed with the module General Linear Model (P < 0.01) and Principal Component and Classification Analysis of Statistica 7.1 based on correlation, and relevant pair correlations were tested with Pearson's r for an additional quantification of the correlations.

RESULTS

Pesticide persistence.

The effect of composting age on pesticide persistence was general, both for single applications and for coapplication, and the amount of residue measured after 125 days was higher in the 12 M compost; this effect was greater for single applications and in particular for AZO, which had a residue in 12 M that was three times that in 3 M. The degradation rate was very low for all pesticides in 12 M compost, and the most persistent pesticide was EPO, with a residue after 125 days of about 70% of the applied dose. Experimental values fitted well with the first-order kinetic degradation model (r² = 0.97 to 0.99), and the calculated half-lives were precise. The longest half-life was for EPO (334 to 513 days), and the shortest was for AZO (116 to 389 days) (Table 2).

TABLE 2.

Persistence parameters of the three pesticides: mean residual amount measured 125 days after application and half-life values calculated with a first-order kinetic

Pesticide	Residual amt measured after 125 days (% of applied dose)								Calculated half-life (days)
	Single application				Coapplication				Single application						Coapplication
	3 M		12 M		3 M		12 M		3 M			12 M			3 M			12 M
	Mean	SE	Mean	SE	Mean	SE	Mean	SE	Value	SE	r2	Value	SE	r2	Value	SE	r2	Value	SE	r2
Azoxystrobin	22.3	0.53	60.9	1.45	34.2	0.09	62.8	0.86	116	2.45	0.99	353	10.1	0.98	162	3.99	0.99	389	16.3	0.97
Chlorotoluron	36.1	0.33	53.0	0.65	42.2	0.54	52.4	0.65	173	3.03	0.99	280	8.10	0.98	204	3.51	0.99	283	5.87	0.99
Epoxyconazole	57.8	0.66	68.1	0.76	68.1	0.62	70.2	0.15	334	7.54	0.99	445	10.8	0.99	484	15.4	0.98	513	11.0	0.99

General pattern of FAME detection.

Seventy-one FAMEs belonging to 9 groups (saturated, monounsaturated, polyunsaturated, branched, hydroxy, methylated, cyclopropane, mixed, and alcohol) were detected. The percentage of FAMEs across detection was variable, ranging from 0.01 to 61.71%, whereas variability between replicates was very low.

The EL method was generally more sensitive than the MIDI method, providing triple the number of detections (4,488 versus 1,461), and was more sensitive for all groups of FAMEs with the exception of alcohol. In total the EL method provided valid detection of 66 FAMEs, while the MIDI method detected only 35, of which 30 were detected by both methods; therefore, 36 and 5 types of FAMEs were unique to the EL and MIDI methods, respectively. The EL method provided a higher number of detections for 57 FAMEs, and the MIDI method provided a higher number for 9 FAMEs, while the two methods were equivalent in always detecting 5 FAMEs: 16:0, 16:1ω6c/16:1ω7c, 18:1ω9c, 18:2ω6,9c/a18:0, and 19:1ω6c/0.846/19cy.

ANOVA.

With main-effects ANOVA the effect of extraction method is significant for 65 FAMEs out of 71 (not significant for 19:1ω9c/19:1ω11c, a13:0, i19:0, 10:0 3OH, 12:0 3OH, and i17:0 3OH) and for all 9 groups. Specifically, with the EL method the total contents in saturated, monounsaturated, polyunsaturated, branched, methylated, and mixed FAMEs are significantly higher than those with the MIDI method, which in turn amplified the content of hydroxy, cyclopropane, and alcohol FAMEs (Fig. 1). At P = 0.05 only the i19:0 content would not differ significantly.

FIG. 1.

Relative contents of the 9 groups of FAMEs, grouped by main factors. Groups of FAMEs marked with a star are significantly different from others. Pesticide abbreviations: UNT, untreated; AZO, azoxystrobin; CHL, chlorotoluron; EPO, epoxyconazole; ACE, coapplication. Composting ages, 3 and 12 months. Degradation times, 0, 56, and 125 days. FAME groups in the shading key are shown according to their positions in the bars in the graphs.

For each extraction method a main-effects ANOVA was performed separately, on 66 FAMEs for the EL method and 35 FAMEs for the MIDI method (see the supplemental material).

With the EL method the effect of pesticide inoculation is the lowest, being significant for only 13 FAMEs. The effect of pesticide is significant for only two groups of FAMEs, the polyunsaturated and cyclopropane groups; for example, when CHL was added, the content of the former increased while that of the latter decreased. With the MIDI method, the lower number of detections obtained profoundly influenced the ANOVA results. In fact, the effect of pesticide is completely absent.

The effect of composting age is of greater importance, and significant differences in the contents of 36 FAMEs were found with the EL method. The composting age had no effect on saturated, polyunsaturated, hydroxy, or alcohol FAMEs, while the content of monounsaturated FAMEs increased and that of branched, methylated, cyclopropane, and mixed FAMEs decreased in relation to compost age. The effects measured by the MIDI method were of less importance, being significant for 19 FAMEs. Taking into account the 9 groups of FAMEs extracted using the MIDI method, the composting age effect is significant, with the contents of saturated and monounsaturated FAMEs significantly increasing and those of branched, hydroxy, and methylated FAMEs significantly decreasing with age.

The effects of degradation time are highly important, and significant differences in the contents of 47 FAMEs were found with the EL method. This factor affected the contents of 5 groups of FAMEs: the contents of mixed FAMEs decreased regularly as time increased, while those of saturated, branched, and methylated FAMEs fluctuated, being lower when time was intermediate or in contrast higher as for monounsaturated FAMEs. The effects recorded with the MIDI method are of less importance, these being significant for 10 FAMEs.

Biomarkers.

The iso and anteiso branched FAMEs, biomarkers for Gram-positive bacteria, were not affected by pesticide inoculation but significantly decreased from 3 to 12 months of compost age; the EL method also highlighted the effect of degradation time, with these FAMEs being 16 to 19% of the total.

The 18:2ω6,9c FAME, a biomarker for fungi, was always detected with both methods, with a mean proportion of 2.26% and 5.16% of the total amount of FAMEs for the MIDI and EL methods, respectively.

The 10Me18:0 FAME, a biomarker for actinomycetes, was affected by fungicide inoculation, and a significant decrease was observed from age 3 months to age 12 months (EL method).

The 16:1ω5c FAME, a biomarker for mycorrhizae, was always detected with the EL method (90 detections) and only seldom with the MIDI method (8 detections). It was quite scarce, with an average content of about 1% of the total FAMEs, decreasing significantly with composting age for both extraction methods. A regular positive effect of the degradation time was detected only by the EL method.

The 20:4ω6,9,12,15c FAME, a biomarker for protozoa, was frequently detected by the EL method (85 detections) and only seldom with the MIDI method (10 detections). Focusing on the EL method, the average content was about 1% and a pesticide effect was found, i.e., the content was lower when the mixture was applied (0.59%) and higher when chlorotoluron was applied (1.23%).

Correlation and classification analysis.

In the principal component and classification analysis the variance displayed by the first two factors is high (53 to 59%) and shows that correlations between contents in the 9 FAME groups generally differed for the two extraction methods.

For both extraction methods the cases classified according to the factor pesticide were randomly scattered in the factorial plane, in agreement with the results of ANOVA that pesticide inoculation had little influence on the FAME profiles, and so here the biplots focus on clustering due to composting age and degradation time classification.

With the EL method the saturated content was clearly inversely correlated with that of monounsaturated (r = −0.35, P < 0.01) and mixed (r = −0.75, P < 0.01) FAMEs but directly correlated with that of methylated (r = 0.59, P < 0.01) FAMEs. For the first factor the fatty acids with the highest correlations were 10Me16:0, 16:0, 18:2ω6,9c/a18:0, and 18:1ω9c; for the second factor they were 18:1ω6c/18:1ω7c, a15:0, and 18:0. A clear inverse correlation was calculated between monounsaturated and polyunsaturated (r = −0.38, P < 0.01) FAMEs (Fig. 2). Detections on 3 M compost at the first two degradation stages (0 and 56 days) formed a cluster related to methylated and branched FAMEs. The cyclopropane content was characteristic of 12 M compost. It is interesting that detections for 3 M at 125 days of degradation formed a common cluster with those for 12 M before degradation began.

FIG. 2.

Biplots after principal component and classification analysis of the contents (%) of 9 groups of FAMEs (variables, ⊕) for compost at 2 different maturity stages and 3 degradation times (case class) detected with EL and MIDI extraction methods. For each group, the position and name of the FAME with the highest correlation with principal components (supplementary variables, +) are indicated. Filled symbols are for compost aged 3 months (3 M), and open symbols are for compost aged 12 months (12 M), both at three degradation times (0 days [circles], 56 days [squares], and 125 days [triangles]). The variance (%) displayed by each principal component is reported on the axes.

With the MIDI method the correlations in content among the FAME groups were particularly simple: on the one hand there was a clear negative correlation between cyclopropane and branched (r = −0.80, P < 0.01) FAMEs, and on the other there was an analogous negative correlation between monounsaturated and methylated (r = −0.51, P < 0.01) FAMEs. As with the EL method, monounsaturated FAMEs were inversely correlated with polyunsaturated FAMEs (r = −0.27, P < 0.01) especially for compost 3 M at time zero (before degradation began). In contrast to the EL method, the cyclopropane content was related to compost 12 M. For the first factor the fatty acids with the highest correlations were a15:0 and 19:1ω6c/0.846/19cy, and for the second factor they were 10Me19:0, 16:0, 18:0, and 18:1ω9c.

Correlation analysis between biomarkers (Fig. 3) shows a clear opposition to Gram-negative bacteria, characteristic of a 12 M compost at all stages, and Gram-positive bacteria, actinomycetes, fungi, and mycorrhizae better related to a 3 M compost at the first two degradation stages, while protozoa show an intermediate behavior. A degradation time effect is apparent for Gram-negative bacteria, and cases of “3 M, 125 days” cluster with “12 M, all stages.” The particular correlation between Gram-negative and Gram-positive bacteria would hold even if the sum of monounsaturated and cyclopropane FAMEs was taken as a Gram-negative bacterial biomarker (19), since the first group is far more abundant than the second.

FIG. 3.

Biplot after principal component and classification analysis of the contents (%) of 6 biomarkers of FAME (variables, ⊕) for compost at 2 different maturity stages and 3 degradation times (case class) detected with the EL extraction method. Gr⁺, Gram-positive bacteria (branched fatty acids); Gr⁻, Gram-negative bacteria (monounsaturated fatty acids); Fun, fungi (18:2ω6,9c); Act, actinomycetes (10Me18:0); Myc, mycorrhizae (16:1ω5c); Pro, protozoa (20:4ω6,9,12,15c). Filled symbols are for compost aged 3 months (3 M), and open symbols are for compost aged 12 months (12 M), both at three degradation times (0 days [circles], 56 days [squares], and 125 days [triangles]). The variance (%) displayed by each principal component is reported on the axes.

DISCUSSION

We evaluated the effect of pesticide inoculation, composting age, and degradation time on the content of compost fatty acids quantified using two methods. Not all effects were clear. It is known that both the MIDI and EL extraction methods appear to be suitable for analysis of microbial FAME profiles in compost (32), but the results of this study show that the extraction method can indeed have a significant effect on the FAME relative content: the MIDI method amplified the importance of cyclopropane, and for all treatments the FAME profile is dominated by 19:1ω6c/0.846/19cy, while for the EL method the profile is dominated by the other main FAME groups, and nearly half of the FAMEs detected consisted of the pair 16:0 and 18:1ω9c, these latter two in proportions in full agreement with those previously reported (25). In general the EL method provided a higher number of detections and a higher number of unique FAMEs and therefore a more equilibrated profile with respect to the MIDI method. This is consistent with the fact that the MIDI method is not specifically designed to generate the maximum number of FAME detections but in general is more attuned to taxonomic identification, i.e., microbial communities from groundwater (11).

An explanation for the differences in the detections may be that with the MIDI method both the 100°C saponification step and the following incubation at 80°C lead to losses of FAMEs (21, 25). This result, together with practical considerations (25), suggests that the EL method has advantages compared with the MIDI method, and so the effect of any treatment on compost should be analyzed with this method.

It is highly probable that the month-based time scale used in this study involves somewhat different processes and modifications of the microbial populations than those typical of the initial stage of composting, i.e., 10Me16:0 (detected only with the EL method) did not change from month 3 to month 12, while it steadily increased from day 0 to day 47 in a previous similar study (32). An effect due to the origin of compost is likely (13) and could explain these differences. The increase in monounsaturated FAMEs from age 3 to 12 months is instead in agreement with previous results showing an increment in Gram-negative bacterial monounsaturated FAMEs over time in compost with a complex composition (2). Furthermore, with the EL method, the ratios of (17:0cy)/(16:1ω6c/16:1ω7c) and (19:0cyω8c)/(18:1ω6c/18:1ω6c) observed were 0.41 and 0.39 and 0.53 and 0.40 for age 3 and 12 months, respectively, showing that the microorganisms had already passed the log phase but not yet entered the stationary phase (2).

In general the proportions of FAMEs are not easily predictable. This can have various consequences, for example, when compost is considered for its capacity to suppress plant pathogens. In fact, on the one hand the proportions of 17:0, i16:0, i17:0, a17:0, and 10Me18:0 tuberculostearic acid (TBSA) decrease from age 3 to 12 months, and the compost can be considered to become more suppressive over time, but on the other hand the proportions of 18:1ω6c/18:1ω7c, a17:1ω9c, and cyclopropane (all with the EL method) decrease, resulting in the opposite conclusion that the compost becomes more conducive over time. This is consistent with the fact that compost disease suppression is known to be a transient property, supported by various different groups of microorganisms employing a variety of mechanisms (2).

In this study the persistence of the pesticides azoxystrobin, chlorotoluron, and epoxyconazole was high, and after 4 months of degradation the residue ranged from 22 to 70% of the applied amount, depending mostly on the composting age. Coapplication of pesticides generally increased persistence. The compost microbial pool has therefore been shown to be not particularly biologically active against the pesticides used, the half-lives of which range from 4 to 15 months. This is consistent with the literature and confirms the recalcitrance to degradation of the pesticides studied. The inoculation of these pesticides did not influence the FAME profile. The abundance of chemical and biological changes that occur during composting, the interactions with the dose inoculated, and the variability of dissipation rate during composting (18) make it difficult to highlight single effects but also to agree on methods for practical assessment. The type of the recalcitrance process, which can be due either to a particular chemical structure or to a strong association with organic matter (30), is also of importance, together with the fact that the degradation pattern and dynamic of each group of pesticides are peculiar and strain dependent (7, 35). Another possibility is that the dominant communities were quite stable and did not change under the treatments compared.

The composting age changed the FAME profile, but the effect on the groups of FAMEs was related to the extraction method. The results of the present study suggest that the use of the sum content of saturated, branched, and methylated FAMEs as a compost maturity index (15) would not be of general use, as the type of original materials can affect FAME composition.

The degradation time is a relevant factor for many groups of FAMEs and in particular is inversely correlated with the content of mixed FAMEs, while for other groups the effect is unclear. It is known that the contents of some fatty acids fluctuate over time according to the availability of original organic matter and the microbial succession, i.e., the amount of fungal biomass. For example, in a 47-day study (32) a kind of succession of fatty acid content during composting was observed: fatty acids common in eukaryotic cells (i.e., 16:0, 18:2ω6,9c, 18:1ω9c, and 18:0) dominated at the start of composting. Later, bacterial fatty acids increased, i.e., iso and anteiso branched FAMEs from Gram-positive bacteria, monounsaturated FAMEs from Gram-negative bacteria, and methylated (i.e., 10Me16:0 and 10Me18:0) and cyclopropane fatty acids from actinomycetes. Then monounsaturated FAMEs reappeared in a successive mesophilic stage. Except for monounsaturated 18:1ω9c, the content of which increased significantly from age 3 to 12 months of composting, with both the EL and MIDI methods, this regular pattern was not observed in the present study, where from age 3 to 12 months the content of 18:0 and 18:1ω9c increased (with both MIDI and EL methods), that of 16:0 increased (only with the MIDI method), and that of 18:2ω6,9c significantly increased with the MIDI method but decreased with the EL method. Furthermore, iso and anteiso branched and methylated FAMEs decreased with age (especially with the MIDI method), and cyclopropane decreased only with the EL method.

Nonetheless, other theoretical succession evidence is somehow contradictory. 16:1ω5c is a marker for arbuscular mycorrhizal fungi, although mycorrhizae are not likely to be an important component of the microbial community in composts, and in fact in the present study this fatty acid proved to be quite scarce and decreased significantly with time. However, other studies (32) came to opposite conclusions, although the time scales investigated were different.

The use of some FAMEs as biomarkers for the main classes of microorganisms is interesting. In this study the iso and anteiso branched FAMEs, biomarkers for Gram-positive bacteria, are in agreement with the fraction of Gram-positive bacteria generally found in various soils (31). Also the fraction of 18:2ω6,9c, biomarker for fungi, is consistent with those of previous studies (29). For both methods a significant composting age effect was detected, but with an opposite trend: with the MIDI method the content of this biomarker increased from age 3 to 12 months but with the EL method it decreased, highlighting again the importance of the extraction method. In other studies content changes of this FAME over a 3-week period were inconsistent (24).

It should nevertheless be acknowledged that the results from this kind of study can be influenced by the type of compost (23) and by the possible fluctuations in the abundance of microorganisms, thus making it difficult to highlight, for example, a clear time effect (5), as observed in the present study for bacterial FAMEs. A certain interference can be due to temperature (21), even if its effects on community level are not always clear (25) because temperature generally fluctuates during composting and is also greatly affected by mechanical mixing (1).

The present study raises issues with regard to putative succession and processes in composting and highlights the need for specific experiments to evaluate the effect of the various sources of organic matter on the microbial succession and FAME profile and the time effect, together with a critical evaluation of the methods based on the FAME profile.