Soil Microbial Communities Involved in Proteolysis and Sulfate-Ester Hydrolysis Are More Influenced by Interannual Variability than by Crop Sequence

Round 1

Reviewer 1 Report

The work done is appreciated and is presented in a good scientific manner. However, the manuscript needs to be revised in view of the points listed below. 

Factors like plants phenological stage, rhizospheric and non-rhizospheric regions and soil depth may have significant influence on microbial abundance and activity. However, the crop phenological stage was not considered in the study.

Microbial abundance and enzymatic activity show significant variation in rhizospheric and non-rhizospheric regions which can influence the out-comes of the present study. Moreover, collecting soil from a depth of 0-10cm may not give an exact idea about the influence of crop as few of the selected crops may have even deeper roots.

Effects of precipitation on microbial activity needs to be discussed at length as it was the major factor influencing the microbial communities.  

Author Response

Response to Reviewer 1

The work done is appreciated and is presented in a good scientific manner. However, the manuscript needs to be revised in view of the points listed below. 

Point 1 Factors like plants phenological stage, rhizospheric and non-rhizospheric regions and soil depth may have significant influence on microbial abundance and activity. However, the crop phenological stage was not considered in the study.

Microbial abundance and enzymatic activity show significant variation in rhizospheric and non-rhizospheric regions which can influence the out-comes of the present study. Moreover, collecting soil from a depth of 0-10cm may not give an exact idea about the influence of crop as few of the selected crops may have even deeper roots.

We added elements to the discussion to mention those limitations.

In l. 387, we added:

“Contrary to other studies, here we studied the whole soil (i.e. bulk+rhizosphere soil), which might impact the results as microbial activities can be greatly enhanced in the rhizosphere compared to bulk soil (Knauff et al., 2003).”

We agree with the importance of phenology. In fact, we think that our experiment design allows disentangling direct effect of seasonality from the impact of plant phenology, as different crop species were at different phenological stages in the same sampling date. WE interpret the absence of interactions between month and crop sequence as an evidence of no effect of crop phenology.

In l.431, we added:

“However, different crop species were at different phenological stage in May and no interactions were observed between crop sequence and arylsulfatase activity.”

We agree that crops can have rather deep root systems. However, the studied crops have of their root biomass in the first 10-15cm of the soil (Fan et al., 2016). Furthermore, due to highly different conditions between shallow in deep soils (akin to a ruderal-stressful gradient of environmental conditions), we do not expect deep soil communities to react in the same way as superficial soil communities. Although we agree that reaction of deep soil microbial communities to crop presence warrant investigation, we think this is out of the scope of our study.

Point 2 Effects of precipitation on microbial activity needs to be discussed at length as it was the major factor influencing the microbial communities. 

We added elements in the discussion on the effect of precipitations.

The paragraph now reads as:

“4.3. Yearly meteorology more strongly impact microbial communities than crops

As hypothesized, the sampling year had a major effect on microbial abundance and activity, which largely overcame the effects of crop sequences and seasonal variability. This effect was likely due to interannual meteorological variability. While the strong influence of climate on microbial abundance has already been demonstrated along spatial climatic gradients [97–99], our study emphasizes the existence of this phenomenon across time. Despite a limited dataset (4 years), we observed a strong impact of precipitations on most of the measured microbial parameters and a more limited impact of temperature. Contrary to most other studies (e.g., Sandor et al., 2011; Zhang et al., 2013; Serna-Chavez et al., 2013) we observed a negative correlations between microbial abundances and precipitations. This might be due to anoxia [103] caused by water saturation in our soil, which had 60% clay content, or by increased nematode grazing at high moisture content [102]. However, positive effects of precipitations on the relative abundance of the studied microbial functional groups show that not all bacterial groups are affected in the same way by meteorological conditions. A positive effect of soil moisture on the abundance of bacteria with arylsulfatase activity was already observed in the same soil (Goux et al., 2012) but the reason for this differential response remains elusive.

One particular case was arylsulfatase activity, for which the sampling month explained half the variation and interannual variability had a limited effect. Because bacterial abundance did not explain arylsulfatase activity in this study, monthly variations were likely caused by changes in arylsulfatase expression. The mechanisms responsible for such a high seasonal change in arylsulfatase expression could be linked to plant phenology. Arylsulfatase activity is known to increase during plant development [22,105], which could result from changes in plant rhizodeposition and S uptake as arylsulfatase is synthesized in response to the limitation of S and to the presence of sulfate ester [77–79]. However, different crop species were at different phenological stage in May and no interactions were observed between crop sequence and arylsulfatase activity. The positive correlations between arylsulfatase activity and both temperature and precipitation could also point to increased enzyme synthesis caused by increased substrate diffusion rate or microbial sulfate uptake, as arylsulfatase synthesis is stimulated by its substrate and repressed by its product (Beil et al., 1995; Cregut et al., 2013).

For most microbial variables, the month and year effects were additive to the crop sequence effect. However, for the three microbial parameters, namely, abundance of fungi, relative abundance of subtilisin and relative abundance of culturable bacteria with arylsulfatase activity, the year and month parameters interacted with the crop sequence, which might mean that year and month effects are stronger under specific crops. This could be due to different crop species responding in different ways to meteorological conditions, for parameters such as phenology, carbon allocation [100], rhizodeposition [57], fine root distribution patterns [106] or nutrient uptake [54–56,107]. This was particularly true for 18S rDNA which was higher in summer than spring under pea, a pattern opposite to all other crops.Further investigation, with longer time series, would be necessary to disentangle those interactions. This is particularly true for the effect of precipitation which display a high temporal variability. It varies both in frequency and intensity within a year, display interannual cycles of variations and present long-term trends of change caused by climate change (e.g., Afzal et al., 2015). As such precipitation might be responsible for high temporal variations of soil microbial communities and warrant further investigations. Unfortunately, agronomic and ecological experiments are rarely replicated across years, despite treatment by year interactions being frequently observed [108]. Our results stress the importance of integrating interannual meteorological variability when investigating plant and cropping practice impacts on microbial communities.”

Reviewer 2 Report

Comments to “Soil microbial communities involved in proteolysis and sulfate-ester hydrolysis are more influenced by interannual variability than by crop sequence”

Using a field trial, the authors have investigated the effect of 6 crop sequences on microbial abundance, enzymatic activity and net mineralization, and they further assessed whether the sampling date effect was stronger under specific crop sequences. They found that  cropping sequences impacted soil microbial communities involved in proteolysis but not those involved in sulfate-ester hydrolysis. They also found that oilseed rape following wheat presented higher relative abundance of alkaline metalloprotease genes and higher protease activity than other cropping sequences. In my view, these findings are timely and important, albeit not entirely novel. The manuscript is well-written throughout and easy to follow. Thus, I think this manuscript should be published after revised.

Specific comments

Line 40, 46: What’s the meaning of “EC 3.4” “EC 3.1.6.1”? 

Lin70-70: A main weak point is the poor integration with previous similar studies. Thus, I advice the authors to take more space and time to introduce the similar studies. The authors should organize the information, and highlight the necessity of the research and the key contributions of the former studies.

Line 115-117: I found that only oilseed rape was systematically fertilized with mineral N and S, but other plots related to wheat, barley and pea received N or S fertilization, or none. I am hesitant whether the different fertilization managements may impact the current results.

Figure 1: There are too many pictures in figure 1. For clarity, split figure 1 into several figures.

In figure S1, 3 or 4 blocks? Please check it.

Line 259: Too wordy. It’s better to only show the mean values.

Line 268-270, 396-397: Again, I am still hesitant whether the different fertilization managements may impact the current results. I advice the authors to take more space to discuss the effects of different fertilization managements on current results.

Author Response

Response to Reviewer 2

Specific comments

Line 40, 46: What’s the meaning of “EC 3.4” “EC 3.1.6.1”? 

EC stands for Enzyme Commission number. This was explicated in the text.

Lin70-70: A main weak point is the poor integration with previous similar studies. Thus, I advice the authors to take more space and time to introduce the similar studies. The authors should organize the information, and highlight the necessity of the research and the key contributions of the former studies.

The introduction (l.67-onward) was re-written to add elements on previous results from studies on soil proteases and arylsulfatases. It now reads as:

“Higher arylsulfatase activity and abundance of bacteria producing this enzyme have been found in the rhizosphere of Brassicaceae than in the rhizosphere of Poaceae (Dedourge et al., 2003; Knauff et al., 2003; Vong et al., 2007; Cregut et al., 2009), which has been attributed to both the higher rhizodeposition and the higher S uptake of Brassicaceae. However, opposite results were found by Wyszkowska et al. (2019), and attributed to a higher biomass production by wheat than by oilseed rape. In one of the rare studies investigating the impact of different crop species on protease activity, Kwiatkowski et al. (2020) found higher protease activity under sugar beet and red clover than under Poaceae, which was attributed to higher root biomass production. In general, protease activity is positively correlated to several soil parameters that are affected by crop species, such as microbial biomass and abundance (Asmar et al., 1992; Geisseler et al., 2012; Giacometti et al., 2013), protein content (Gispert et al., 2013) and nitrogen availability (Tian et al., 2013).

Nevertheless, knowledge gaps remain regarding the effect of crop identity and crop succession on soil microbial communities involved in N and S cycling and the resulting net mineralization fluxes. Filling those gaps is necessary to design cropping systems in which soil fertility management is improved and adapted to the pedoclimatic context [47,48].

Seasonal and interannual variability in temperature and precipitation can also strongly influence soil microbial communities’ abundance, structure and activity [49–53]. The soil water content is positively correlated to the microbial biomass (Serna-Chavez et al., 2013) and influence the bacterial community composition (Lauber et al., 2009). Soil water content control the diffusion rate and the concentration of enzyme’s substrates and products, thus modifying the enzymatic reaction’s rates and the perception of the environment by the microorganisms (Sinsabaugh et al., 2008; Allison et al., 2011; Nunan 2017). Temperature impacts similarly enzyme’s reaction rates and organisms’ growth rates (Allison et al., 2011). Indirect effects can also occur through increased nematode grazing in wetter soils(Gray et al., 2011) and modification of plant growth and rhizodeposition. For example, Lauber et al. [57] hypothesized that seasonal variations in bacterial community composition in agricultural soils could be related to changes in plant C inputs to the soil (i.e., root exudates and plant litter) caused by changes in soil temperature and moisture over the season. Positive effect of soil moisture and temperature have been observed on protease (Mancinelli et al., 2013; Fraser et al., 2013) and arylsulfatase activity (Vanegas et al., 2013), even if mechanisms responsible for it remain unclear (Bell and Henry 2011). It is thus important to assess the impact of cropping sequences on soil microbial communities over multiple years to account for seasonal and interannual variations.

As field experiments are complicated the by high seasonal and interannual variability of microbial variables, plant impact on microbial communities have been evaluated more often in microcosms than field conditions, (Schjonning et al., 2007; Soon et al., 2007; Yin et al., 2010; Vaughn et al., 2010; Liu et al., 2010; Reardon et al., 2014). As a consequence,  the relative importance of climate and crop effects on microbial communities, as well as their potential interactions [58], remain largely unresolved and no study to our knowledge compared the influence of climatic conditions and different crop species on microbial communities implicated in protease and arylsulfatase activity.”

Line 115-117: I found that only oilseed rape was systematically fertilized with mineral N and S, but other plots related to wheat, barley and pea received N or S fertilization, or none. I am hesitant whether the different fertilization managements may impact the current results.

We agree with the reviewer that fertilization probably impact microbial communities, at least indirectly by stimulating plant growth. However, we wished to compare crops in realistic situations and no crop comes without its fertilization regime. Here we stuck to local fertilization practices. We consider our results valid when comparing crops under conventional cropping in Northern France, while they might be different under organic farming our in other socio-economic and cultural settings. Anyways, we think there would be no point in comparing crops with identical fertilizationn as those would be unrealistic situations. We added “and are representative of local practices » in l.112.

Figure 1: There are too many pictures in figure 1. For clarity, split figure 1 into several figures.

Figure 1 was split in 2 figures.

In figure S1, 3 or 4 blocks? Please check it.

Four blocks, this was corrected.

Line 259: Too wordy. It’s better to only show the mean values.

This part was simplified. The sentence l. 266-269 was deleted

Line 268-270, 396-397: Again, I am still hesitant whether the different fertilization managements may impact the current results. I advice the authors to take more space to discuss the effects of different fertilization managements on current results.

See answer to previous comment. We added in l.387:

“Other differences between studies can comes from differences in fertilization regimes. In particular, organic fertilization is known to increase microbial abundance (Kallenbach and Grandy 2011), protease (Giacometti et al., 2013) and arylsulfatase activities (Knauffe t al., 2003) when compared to mineral fertilization.”