2.2 Proteomics in Synechocystis
2.2.3 Standard procedures in use and protein identification challenges
lenges
There is currently no community-standard method for generating proteomic data from
Synechocystis, however there are a number of techniques in the cell-pellet processing
pipeline that follow broadly the same patterns. Cell disruption is the process of lysing the cells. There are two techniques that appear in the majority of publications reporting high protein identifications: sonication and bead-beating. These have become popular because they are automated, effective and highly reproducible. Other traditional techniques, such as liquid nitrogen grinding, have a great variation between different users and the methods are difficult to report accurately and reproduce. High-pressure methods, such as the French press or Yeda press, have been shown to be thoroughly ineffective at disrupting
Synechocystis cells, as certain strains are pressure resistant. More accurate comparisons
between extraction methods is impossible as no study has directly compared extraction methods whilst maintaining the same conditions downstream.
2.2. PROTEOMICS IN SYNECHOCYSTIS 71 Cell debris processing is where the majority of divergence in publications occurs. This refers to the process of taking the disrupted cellular material and processing it into peptides for mass spectrometry analysis. Broadly there are two methods of doing this, separating the whole proteins out on a gel and using in-gel digestion to produce peptides or digesting individual samples in solution and fractionating the peptides out by their individual features. The method chosen is heavily dependent on the down-stream pro- cessing and the focus of the experiment being carried out – if the user is conducting an exploratory investigation that does not require merged quantification methods then the protein-level separation has been shown to have significant advantages over peptide-level separation. A more in-depth comparison between these methods is made later in this review.
Peptide post-processing is an optional step, included in studies where the user is inter- ested in identifying post-translational modifications or applying a peptide-specific tag. Post-translational modification purification is a rapidly growing topic which requires a specialised approach for each different modification being concentrated. For further de- tails on current trends in PTM research methods, please see this review by (Huang et al., 2014). With over 20% of all Synechocystis proteomics papers citing their use, the most commonly used peptide-specific tags are iTRAQ (isobaric tags for relative and absolute quantification). Alternative quantification tags have only rarely been used in Synecho-
cystis with a single study reporting the use of TMT (tandem mass tags). To date no
clear comparison between iTRAQ and TMT tags has been made.
Quantification of proteins in proteomics is generally done in one of two ways, either through a gel-based method using 2D gel electrophoresis stain intensity analysis, or else through a spectral intensity counting method on the mass spectrometer. The quanti- fication method used is dependent on the mass spectrometer available; gel methods are commonly associated with MALDI spectrometers whilst spectral counts are used with the LC-MSMS setup. Almost all studies to date have processed spectra for identifica- tion with the software Mascot (www.matrixscience.com). Alternative programs have been used previously, but a flat comparison between different software has not been completed. Proteomic techniques in Synechocystis have been rapidly evolving over the last 5 years. This can be seen broadly from the number of protein identifications that have been reported per publication (Fig 2.1 p. 72). A severe limitation on protein identification in Synechocystis is the presence of small, high-abundance phycobili-proteins (Gan et al., 2005). These proteins are integral to the light harvesting antennae, which account for around 40% of all proteins in the cell or 20% by weight due to their relatively small size – which can be observed on a standard protein poly acrylamide gel electrophoresis (PAGE) analysis as shown in figure 2.2 (p. 72). The focus of these studies has also been changing with time, ranging from initial investigations into the membrane eventually leading to
Figure 2.1: The number of proteins identified in each proteomic study of Synechocystis per year. All studies that confidently identified more that 1000 proteins are highlighted in yellow, all the green points were conducted by the same lab over the last 4 years, and the point in red was the first study that focused primarily on increasing the number of protein identifications.
Figure 2.2: A slice of a gel image, showing a size ladder in the top row and Synechocystis proteins in the bottom row. The four strongest bands on the gel are the phycobiliproteins, which dominate the protein sample. The blue colour on the proteins here is as a result of dyeing with the Bradford reagent, however the phycobiliproteins also showed as a blue shift on the band whilst the gel was running.
much broader-reaching studies investigating technical improvements to techniques and systematic development of the organism for biofuel production (figure 2.3 p. 73). There have been several limits that have been explored in proteomic investigations, as detailed in table 2.2 (pg. 74).
2.2.4
Synechocystis proteomic studies
To date, only around 65% of the proteome of Synechocystis has been successfully identified through mass spectrometry, although this is higher than the number of annotated genes in the genome.
2.2. PROTEOMICS IN SYNECHOCYSTIS 73
Figure 2.3: The topics being published in Synechocystis over time. Each publication was given a tag based on the topic, which are as follows: mut – Mutant study; rev – Review; starv – Stress, Starvation; tol – Stress, Tolerence; tech – Technical improvement study; exp – Exploritory Studies. In cases where a publication addresses multiple topics, it was assigned multiple tags, to reflect current knowledge and direction of interest in the field. No single publication was given multiple counts of the same tag, regardless of the size of the study.
Figure 2.4: The overall number of publications per month since January 2012. The levels remained largely steady until a large spike in publications at the beginning of 2015 – the dip at the end of the graph results from an incomplete set of data, as this was collected over the first 16 days of March, 2015; suggesting a continuation in the publication trend. This figure is taken from the supplementary materials of (Landels et al., 2015).
Table 2.2: A table listing the current limits that have been investigated with a proteomic study in Synechocystis. Whilst these data cover a broad base of topics – many of which are discussed in more detail in the following sections of this section – as can be seen from figure 2.1 (pg. 72), a number of these studies conducted before 2010 may be of limited use compared with data that could be obtained with better analysis capabilities.
Condition Limit Temperature Low: 20 degrees, High: 38 degrees Shock: 44 degrees pH Low: 5.5 High: 11 Complete starvation Nitrogen – 6 days Phosphorus – 6 days Sulphur – 6 days Iron – 6 days Longest
study Low phosphorus – 60 days
Light
Highest intensity light: 300 mmol photons/m2.s Highest intensity UV: 1 W/m2
Lowest intensity: Complete Darkness Biofuels 0.25% Butanol 2% Ethanol 0.9% Hexane Salt High: 6% NaCl (9 days) {CO2} High: 3% CO2
Low: air level Metal Toxicity Cadmium 40 µM Cobolt 40 µM Nickel 40 µM Metal Tolerence Cadmium 40 µM
2.2. PROTEOMICS IN SYNECHOCYSTIS 75 we highlight techniques and features of proteome investigation that yielded the highest degree of success. The single point in red is the first investigation into improving protein identification in Synechocystis, performed by (Gan et al., 2005). The authors highlighted the importance of separating out the membrane and soluble fractions of the proteome prior to digestion, to enable solubilisation and digestion of the membrane fraction before recombining the two fractions together. Here a total of 776 proteins were positively identified with isoelectric focusing for protein and peptide fractionation prior to reverse phase giving the best overall number of positive proteins identifications (Gan et al., 2005). The technique is now a standard in this field.
In Figure 2.1 (p. 72), the points highlighted in green all reference the same protein extraction protocol from the Zhang group (Qiao et al., 2012a), and are produced in the same lab using the AB SCIEX Triple ToF 5600 spectrometer. These publications provide both qualitative and quantitative data through the use of the isobaric tagging agent iTRAQ. The most notable difference between these studies and others in the field with far fewer positive protein identifications appears to be the use of the high quality mass spectrometer. Whilst this is clearly an effective solution to the problem, it is unhelpful to labs that lack the resources required to purchase more expensive, powerful machines. The orange points are all other published studies that have produced more than 900
unique protein identifications. The earliest Synechocystis publication to exceed this
threshold was a bioinformatic-based protein identification investigation (Ishino et al 2007). This study was notable as instead of using offline LC separation, proteins were size selected using electrophoretic mobility and in-gel digestion before analysis with reverse- phase LC-MSMS. This qualitative method is affordable and highly effective – reporting 1442 unique proteins with good confidence but has yet to be used successfully with quan- tification methods, such as isobaric tagging.
In 2010 Wegener et al identified almost 2000 proteins ((Wegener et al., 2010). This study combines the output of 12 different conditions, each double-injected into the mass spec- trometer. This dramatically increases the number of total identified proteins, suggesting that this comprehensive method of obtaining data from samples isolated from a variety of conditions is sufficient to increase the level of proteome coverage at the expense of time. The three remaining highlighted studies that exceed the selection criteria utilise depletion methods; one at the cell-pellet extraction level where phycobili-proteins are washed out of the membrane fraction (Zhang et al., 2013a) and the other two through post-digestion peptide purification – in this case a side-effect of preferentially selecting cys-containing peptides (Guo et al., 2014) or phospho-peptides (Talamantes et al., 2014).
There is evidence to suggest that the reduced protein identification is a stochastic feature. The study by Wegener et al doubled the number of unique protein IDs that had been made
at that time whilst identifying the majority of already identified proteins (Wegener et al., 2010). Notably, their work combined data from a number of different stress conditions with data from standard conditions. Low-abundance proteins in Synechocystis can be considered to fall below a ‘stochastic selection threshold’. Whilst it is feasibly possible to identify a unique peptide from one of these proteins, it is highly unlikely that the second unique peptide required for confident identification would be found. In the Wegener study, stress proteins that would normally be of low abundance became much more prevalent, therefore enabling identification. By running multiple comparisons and producing much more data, they lowered the ‘stochastic selection threshold’ for the peptides generating large numbers of identifications.