What are the maximum levels of salts and detergents that can be in my solution samples?
Quality of results depends heavily on the contamination load of the samples. Ensure samples contaminants are below the values listed:
|| Maximum concentration
| Alkali metal salts
| Tris buffer
| Ammonium bicarbonate
| Phosphate buffer
| All surfactants/detergents
So, silver staining CAN be used to visualize proteins prior to MS analysis?
Only if you must! We don’t recommend it. Not all silver staining protocols are MS-compatible. Subsequent MS limits the choice of reagents which can be used for silver staining because the proteins must not be extensively modified by the procedure. Therefore, common sensitizing reagents such as glutaraldehyde should not be used, and overexposure to silver nitrate should be avoided. If you must use a silver stain to visualize your proteins, however, we recommend using a MS-compatible silver staining protocol (Shevchenko et al., 1996; Yan et al., 2000). You’ll find these required amendments to the silver stain protocol get you sensitivities not that much greater than colloidal Coomassie. Silver-stained gel fragments submitted to the SAMS Centre for MS analysis are destained to remove silver ions before enzymatic digestion to optimise the sensitivity and quality of mass spectra (Gharahdaghi et al., 1999).
What method should I use to visualize proteins in my gel?
Gels are most commonly stained post-electrophoresis using Coomassie, silver, and Sypro fluorescent stain protocols. Commassie (colloidal and regular) and Sypro stains are, by far, the most mass spectrometry friendly. Sypro, being more sensitive is probably the best choice for visualizing low abundance proteins. However, there is still the issue of sample amount. Just because it can be visualized in a Sypro stain does not mean that we will be able to identify it by mass spectrometry. Remember that we prefer to have at least 10 ng of protein for identification attempt. Since these amounts can be just detected by Coomassie, it is still preferred from our perspective.
How should I ship you my gel bands?
Keep the protein contained within the gel slice/band slightly acidified to reduce the activity of proteases. Use 1% acetic acid, and store at 4 degrees C if you can’t ship them to us immediately.
Should I send you my sample frozen or not?
Do not freeze the gel bands/spots. We have found that thawing gel bands that have previously been frozen may result in leakage of intact proteins from the gel matrix during subsequent handling.
How should I cut out my gel band/spot for MS analysis?
Bands or spots of interest from 1-D or 2-D gels should be excised with as little excess gel as possible using a clean instrument (i.e. scalpel blade or spot picker), in a laminar flow hood or other “clean” environment. Keratin is the enemy! Wipe down your excising tools and surfaces with lint-free cloths using 70% ethanol. This includes your gloves and work bench! Put the excised gel piece into a prewashed microfuge tube (0.6ml or 1.5ml tubes are fine) and they are ready to send away.
As an aside, silver nitrate is often contaminated with keratin and is one of the many reasons why we like to avoid silver-stained gels.
I would like to confirm the purity of my peptide synthesis, or the mass of my protein. Can I use MS?
Yes! Submit your sample for MALDI-TOF. For local users, we have a useful walk-up service. We’ll train you how to do it yourself.
I was hoping to see if my protein was modified at a certain site. Why can’t I get this information?
The experiments we perform for protein identification provide only a certain level of “identity”. You obtain a random amount sequence information in most cases – from as low as 2% sequence coverage to as high as 100%. Most typically we are under 50% sequence coverage. In routine analyses, we focus our enrichment, detection and bioinformatics approaches on unmodified hydrophobic peptides with a minimum of modifications (e.g., we avoid glycopeptides). Admittedly this is a bias, but it is designed to maximize the likelihood of a basic identity. If additional sequence-level or modification-level data is required, further custom analyses could be performed upon consultation.
How are the results from an LC-MS/MS different from the fingerprint results? How should I interpret them?
The MS/MS process generates partial sequence data that is far richer than simple peptide molecular weights for obtaining positive protein identification. Peptide ion fragmentation spectra are searched against a database in a fashion similar to the fingerprinting approach, in that a probability of identification is returned. In these types of analyses, we like to see at least two (or more) peptides with statistically significant scores. “One-hit wonders” are possible but the ion score should be very high. So, there is greater identification power in MS/MS data than in fingerprint data; fewer peptides are needed and so more complex protein mixtures can be processed. By combining MS/MS with a chromatographic separation of the mixture digest, you have a rugged tool for sample cleanup/enrichment as well.
For more information on how to interpret your MS/MS data, go to the Mascot website here, or talk to SAMS staff.
I’ve taken your advice and requested an LC-MS/MS analysis, because the MALDI fingerprint approach was not successful. Do I need to give you more sample?
Not usually. We never process the entire digest in MALDI-TOF. We’ll simply use the remaining sample for LC-MS/MS. In fact, we save your processed MALDI sample for one month, in case you come back and request re-analysis via LC-MS/MS.
I submitted my sample for MALDI-TOF analysis but my data didn’t generate good search results.
This isn’t uncommon even with good quality MALDI-TOF data. A good quality peptide fingerprint data-set is one that contains 10 or more peptides in the MALDI-TOF spectrum. Let’s assume you’ve obtained a good spectrum. The usefulness of the dataset depends primarily on two issues:
- The organism of origin
Certain organisms have well-annotated databases, such as C. elegans or E. coli and are inherently less complicated than higher organisms. Peptide mass fingerprints work well in these situations. We achieve success rates in excess of 75% in these cases. However, when analyzing human samples, for example, the success rate drops below 25%, even though the spectral quality is high. Larger databases, incomplete annotation, gene sequencing errors and abundant post-translational modifications combine to diminish the utility of the fingerprint approach.
- The complexity of the protein digest
If your sample contains more proteins than anticipated (very common in 1D gel separations), then a fingerprint will not work very well either. The fingerprinting approach provides a “best-fit” solution. The full list of peptides is assumed to arise from one or two proteins, and the best hits returned from a search are those that best explain the submitted data set. In short, the probability of a good hit diminishes with increased spectral complexity arising from the presence of multiple proteins. This is one reason why contaminating proteins are important to avoid. The other reason is more subtle but equally important: the MS techniques that are applied have a limited capacity for peptides. Just because the sample has more peptide does not mean that more will be detected. The resulting phenomenon we call ion suppression, where only the most abundant peptides and/or the most efficiently generated peptide ions will be detected.
The MALDI TOF gives us insight into these issues, and helps us design the next (and more useful) identification strategy (LC-MS/MS). Equally important, we can halt the analysis if the levels of contamination are obviously very high, and recommend to you that further cleanup steps be applied before analysis.
What type of report do I receive for a peptide mass fingerprint?
The SAMS Centre for Proteomics uses Mascot as a database search tool. For a peptide mass fingerprint, you will be supplied with the raw data, which can be viewed in a program called m/z (download here). You would also be given a search result which would look similar to this. Detailed descriptions of data interpretation can be found on the Mascot website here.
I want to determine the identity of a protein. What MS approach should I use?
Many simple protein samples submitted for identification could be analyzed by MALDI-TOF first, which represents a rapid, inexpensive yet sensitive MS approach. This involves the MALDI-TOF analysis of a tryptic digestion of your sample to generate a ‘peptide mass fingerprint’ (PMF). This ‘fingerprint’ relies on the specificity of trypsin and can be used to search databases of theoretical in silico tryptic protein digests. When database matches are found, a wide range of information on the newly-identified protein automatically becomes available from the database.
If MALDI-TOF analysis proves uninformative (which is not uncommon for samples of mammalian origin) or if protein identification on the basis of sequence data is required, a short LC-MS/MS run may be a more suitable approach. In LC-MS/MS analyses, tryptic peptides are first separated by HPLC, then iteratively mass-selected and fragmented to generate partial sequence information, which can be used to identify a protein from as little as a single peptide. The HPLC part enables the identification of many more than one protein at a time. This may be useful if contaminating proteins in your sample prevent conclusive identification of your protein of interest by MALDI-TOF analysis, or if there is comigration of proteins in your gel separation (e.g. in a 1D gel purification of an immunoprecipitate).
We’d like to note that all identification services that we offer begin with a tryptic digestion. We will always collect a MALDI-TOF spectrum on a small portion of your sample, as a “quality control” step, even if other analytical services are requested. It gives us the opportunity to dialog with you about contamination, for example, that may prevent a successful LC-MS/MS analysis.
How much sample do I need for a successful protein identification?
Basically, if you can see the protein in a Coomassie stain of your gel, we can get a good mass spectral dataset from it (and we do prefer gel-based samples). Our analytical protocols have been validated on fmol amounts of protein using known standards (e.g., ~70 pg of digested BSA) but this doesn’t always translate into similar sensitivities for “real-world” samples. Lower digestion efficiency at low substrate concentrations, incomplete sample extraction and the attendant sample handling losses conspire to reduce overall sensitivity. By providing us with no less than 200 fmol (or 10 ng of a 50 kDa protein) you ensure a maximum probability of success. Avoid using silver-staining techniques for your gel samples – it is easy to overexpose your gels, which hampers processing for analysis. If you are providing your sample as a solution or a lyophilized fraction, try your best to provide at least 50 to 100 ng (the more the better).