Running head: 15N-SIP methods
Keywords: nitrogen, 15N, DNA-SIP, RNA-SIP, amplicon sequencing, BNF, diazotrophs
1 Introduction
1.1 Background on 15N-SIP
Nitrogen is the 3rd most abundant element in living cells by weight and is essential for synthesising proteins and nucleic acids \cite{Fenchel_2012}. Although the atmosphere is composed of nearly 80% N2, nitrogen is biologically unavailable in this form and organisms must therefore either acquire fixed nitrogen forms from the environment or produce them themselves \cite{madigan_chapter_2017}. Some microorganisms, known as diazotrophs, can reduce dinitrogen gas into ammonia for their own needs in a process termed biological nitrogen fixation (BNF). BNF is one of the most energy costly processes in nature, requiring 16--24 moles of ATP for each mole ammonia produced \cite{Fisher_2002}. It is therefore not surprising that BNF is tightly regulated on both transcriptional- and post-translational levels in the cell \cite{Fischer1994,Dixon2004}. This makes studying the ecology of diazotrophs in the natural environment challenging since detecting genes, or even transcripts cannot provide a guarantee for the activity of nitrogen fixation. Studying diazotrophy using stable isotopes probing is, therefore, advantageous because it can provide a strong link between activity and genetic identity.
Despite the cardinal importance of nitrogen to life and the centrality of nitrogen in many ecosystems, only a handful of studies involving stable isotope probing using nitrogen (15N-SIP) have been published to date. There are several good reasons why 15N-SIP is not as nearly as popular as 13C-SIP or even 18O-SIP. Probably the most important one is the fact that while N-transforming processes always drew considerable attention from microbiologists, many of them are dissimilatory, used solely for gaining energy and not for building biomass. Nitrogen assimilation is limited to BNF, or otherwise assimilation of fixed nitrogen forms from the environment such as ammonia, nitrate or amino acids (or peptides). However, assimilation of fixed nitrogen forms is widespread amongst most organisms and therefore provides relatively little differentiating power for targeting specific microbial taxa or guilds in natural communities. A second and obvious reason is that of the three types of biomarkers used for SIP, namely nucleic acids, proteins and lipids, only the first two contain nitrogen and can be used as targets for 15N-SIP, while lipids are excluded. An additional important reason for the relative lack of popularity of 15N-SIP is the difficulty in getting the cells to assimilate enough of the isotopic label. First, because of the high energetic costs, diazotrophs will typically only fix so much atmospheric nitrogen as to fulfil their basic requirements, so a high level of 15N assimilation is difficult to achieve. Secondly, and more important, is the fact that nitrogen atoms are much less abundant than carbon in proteins and nucleic acids, thus inevitably leading to a lower maximum mass addition upon labelling. As a result, while 13C-labelling yields a density gain of ca\(.~\)0.036 and 0.035 g ml-1, 15N-labelling yields only a density shift of 0.016 and 0.015 g ml-1 in fully labelled DNA and RNA, respectively \cite{Lueders2004,birnie_isopycnic_1978,Angel_2017}. Lower density shifts of labelled DNA and RNA mean a greater overlap between labelled and unlabelled templates, which creates a significant challenge for analysing 15N-SIP data. In RNA-SIP, a greater overlap makes it more difficult to detect the enrichment of sequences above the background level. The problem is even more critical in DNA-SIP because DNA also migrates as a function G+C content and could cause unlabelled high-G+C sequences to become enriched in the heavy fractions of the gradient without being labelled.
One successful way to overcome this was published in 2007 and used a two-step centrifugation protocol and the DNA-intercalating agent bis-benzimide \cite{Buckley2007a}. Briefly, the method works as follows: a first density gradient is prepared and centrifuged following a standard DNA-SIP protocol. Then, the heavy fractions corresponding to a density of ca\(.~\)1.725--1.735 mg ml-1 are collected and pooled together. These fractions presumably contain labelled DNA of relatively low G+C content with unlabelled DNA of high G+C content. The DNA in these fractions is then used for a second centrifugation step in a CsCl density gradient containing bis-benzimide. During the second centrifugation step, bis-benzimide significantly decreases the BD of low G+C content DNA thus resolving it from unlabelled high G+C DNA (Fig. 1C). However, more recent works employing 15N-DNA-SIP tended to avoid this two-step protocol and instead rely on the ability of high-throughput sequencing coupled with statistical modelling to detect labelled taxa and avoid false positives via the use of parallel no-label controls \cite{Pepe_Ranney_2015} (see Fig. 1 B and Chapters 9 and 11). The first published attempt at 15N-RNA SIP is attributed to Addison and colleagues in 2010 \citep{Addison2010}, although the authors finally concluded that 15N-labelled RNA could not be definitely resolved from unlabelled RNA. However, it should be noted that the protocol used in that work deviated somewhat from the standard RNA-SIP protocol in several aspects, including using much higher amounts of RNA, higher centrifugation speed but lower temperature and shorter centrifugation time. Finally, a successful demonstration of a 15N-RNA-SIP protocol was published in 2018 and using a standard RNA-SIP protocol in combination with amplicon sequencing and statistical modelling \cite{Angel_2017}.
1.2 Experimental considerations
As in any SIP experiment, incubating the sample in the presence of the 15N-labelled substrate should be prolonged enough to ensure that the DNA or RNA are sufficiently labelled above the detection limit. In contrast, long incubation times will almost inevitably result in the labelling of non-diazotrophic microbes through cross-feeding. The issue of cross-feeding is of general concern in SIP experiments and has been mostly discussed for 13C-based SIP experiments (e.g., \citealt{McDonald_2005,DeRito_2005}), but diazotrophs have also been shown to release substantial amounts of fixed nitrogen through cross-feeding or leaching \cite{Belnap_2001,Adam_2015}. Diazotrophy is a slow and costly process, and, incubation times are accordingly relatively long compared to incubations with a 13C-labelled substrate. Consequently, 15N-SIP incubations targeting diazotrophs would require incubating the samples for several days or even weeks, depending on the specific level of activity of the system \cite{Angel_2017,Buckley_2007,Pepe_Ranney_2015}. However, for targeting the assimilation of biologically available N-forms such as ammonium, nitrate or amino-acids, incubation times should be reduced to several hours to few days, since the process is much more rapid and requires only little energy from the cells \cite{Alonso_Pernas_2017,Bell_2011}. Because of the greater overlap between labelled and unlabelled sequences in 15N-SIP compared to 13C-SIP gradients the chance of detecting false negatives and false positives increases. This can be remediated to some degree by increasing the number of replicates in the experiment.
1.3 Data analysis
In essence, data analysis for 15N-SIP experiments is not different from what is used in other DNA- and RNA-SIP experiments. 15N-SIP experiments were analysed successfully using both traditional comparison of clone libraries \cite{Buckley_2007} and statistical modelling of high-throughput amplicon data (HR-SIP)\cite{Pepe_Ranney_2015,Angel_2017} and qSIP \cite{Morrissey_2018}. The much greater sensitivity achieved through high-throughput sequencing and statistical modelling is particularly advantageous for analysing 15N-SIP experiments, because the target guild is typically small and only partially labelled, and because the separation between labelled and unlabelled nucleic acids is low. Figure \ref{816843} illustrates the results of a 15N-SIP experiment using only two phylotypes that differ in their G+C content and where only the low-G+C-content organism can assimilate 15N. Using standard DNA-SIP procedure the labelled and unlabeled phylotypes cannot be visually differentiated, because of G+C-content based density shift (Fig. \ref{816843} A and B). However, using a second centrifugation step in the presence of bis-benzimide helps to resolve the two phylotypes (Fig. \ref{816843} C). However, the two phylotypes can also be resolved using a standard one-step centrifugation if the results are statistically modelled using qSIP or HR-SIP (Fig \ref{816843} A and B). On the other hand, in RNA-SIP because G+C content has relatively little effect on buoyant density in the presence of formamide, the small mass addition from 15N labelling is visible. Nevertheless, using statistical modelling to detect labelled phylotypes is nevertheless advantageous or even necessary in most real-life cases because of the increased sensitivity.
1.4 Procedures
Methodological details for conducting both DNA-SIP and RNA-SIP are given below. In general, the steps for performing SIP are independent of the isotope used, so the protocols below can be used for processing samples from any DNA- or RNA-SIP experiment. For DNA-SIP, a protocol including a secondary centrifugation step in the presence of bis-benzimide is also detailed as an optional deviation from the standard DNA-SIP protocol. Although this method is considered outdated by now, it is provided here for completeness. Since DNA- and RNA-SIP protocols share many similarities with each other, much of the protocol is given for both methods together, and deviations for each specific method are highlighted. All protocols assume that an environmental sample has been incubated in the presence of a 15N-labelled substrate and that total DNA or RNA have been extracted from the sample following incubation (see Notes 1 and 2). Methods for extracting DNA or RNA from environmental samples are well established and go beyond the scope of this chapter. Many commercial kits are available for this purpose, depending on the type of sample, as well as also general-purpose lab protocols (e.g., \citealt{Angel_2012}).
2 Materials
2.1 Gradient preparation
- An ultracentrifuge, capable of achieving 177,000 \(\times\) g and equipped with a vertical or a fixed-angle rotor for tube volumes of 2--8 ml (typically 5--6 ml; e.g., VTi 90 from Beckman Coulter)
- Compatible polyallomer ultracentrifugation tubes and caps (one for each sample, e.g., Optiseal 4.9 ml)
- Refractometer (typically, Reichert's AR200 digital refractometer)
- DNA samples in TE or water (0.5--5 µg; for DNA SIP)
- CsCl solution (prepare a 7.163M CsCl solution by dissolving 603 g CsCl in 500 ml of filter-sterilised molecular-grade water; confirm that the density is ca\(.~\)1.89 g ml-1; store at RT; for DNA SIP)
- RNA samples in TE or water (300--500 ng; for RNA SIP)
- CsTFA solution (ca\(.~\)2 g ml-1; store at 4 °C; for RNA-SIP)
- Hi-Di™ Formamide (Thermo), or any other deionised formamide (for RNA-SIP)
- Gradient Buffer (GB): prepare a 0.1 M Tris-HCl (pH 8.0), 0.1 M KCl and 1 mM EDTA in RNase-free water, filter-sterilise (0.1 µm) into a clean glassware and autoclave
- Gradient Buffer (GB): prepare a 0.1 M Tris-HCl (pH 8.0), 0.1 M KCl and 1 mM EDTA in RNase-free water, filter-sterilise (0.1 µm) into a clean glassware and autoclave
- One 50-ml tube per gradient
- RNase-free water (for calibrating the refractometer)
- Bis-benzimide (Hoechst 33258, 10 mg ml-1 solution; for DNA-SIP using bis-benzimide)
2.2 Gradient fractionation
- Refractometer
- 1.5-ml non-stick tubes
- Test tube utility clamp mounted on a stand
- 20-ml syringe
- A flexible tube (approx. 30 cm; for instance an elastic HPLC tube) attached to the syringe on one end with a Luer-Lock connection fitting, and with an additional Luer-Lock connection fitting for a disposable needle on the other end
- RNase-free water for displacing the gradient solution (enough to displace the entire volume of an ultracentrifugation tube times the number of gradients)
- Variable-speed, automatic syringe pump
- Disposable needles: 23G and 26G
2.3 DNA-SIP fraction precipitation
- GlycoBlue™ Coprecipitant (15 mg ml-1) or molecular-grade glycogen (see Note 5)
- PEG 6000 solution (prepare a 30% PEG and 1.6 M NaCl solution by dissolving 150 g of polyethylene glycol 6000 and 46.8 g NaCl in molecular-grade water to a final volume of 500 ml and autoclave. Final solution is 30% PEG 6000 and 1.6 M NaCl)
- Ethanol (prepare a 70% solution using molecular grade ethanol and molecular-grade water)
2.4 RNA-SIP fraction precipitation
- GlycoBlue™ Coprecipitant (15 mg ml-1) or RNA-grade glycogen (see Note 5)
- Ethanol (100%; molecular grade)
- Sodium acetate solution (3 M; pH 5.2, RNase free)
- Ethanol (70%; molecular grade in RNase-free water)
- Optional: RNA Storage Solution (Ambion)
3 Methods
3.1 Gradient preparation and centrifugation
- Prepare all solutions in advance.
- Equilibrate the CsCl (for DNA-SIP) or CsTFA (for RNA-SIP) solution to room temperature for about 60 min (if stored at 4 °C).
- Calibrate the refractometer using pure water.
- Prepare the gradient mixture depending on the type of SIP (see below):
3.1.1 Preparation of DNA-SIP gradient mixture
- For each gradient, mix GB, DNA sample and CsCl solution to reach the desired density (typically 1.725 g ml-1) in a separate 50-ml tube. The volume of CsCl solution needed to achieve a specific density is given according to equation \ref{eqn:mix1}.