Distinguishing the effects of Isoniazid and Rifampin on Mycobacterium tuberculosis: A comparison of in vitro and in vivo proteomics Nicole A. Kruh1, JoLynn Troudt1, Angelo Izzo1, Jessica Prenni2 and Karen Dobos1 Department of Microbiology, Immunology and Pathology1; Proteomics and Metabolomics Facility2, Colorado State University, Fort Collins, CO Abstract In vitro Scaffold Results & Spectral Counting Multiple Reaction Monitoring (MRM) For nearly half a century, a combinatorial chemotherapy approach has been taken for Initial proteomic data was collected as full-scan MS/MS spectra in a data-dependent treating Mycobacterium tuberculosis (Mtb). While the treatment of pan-sensitive strains is a The program Scaffold was used to compile all database search results. This program allows us well-established six-month regimen, the treatment of multi-drug resistant tuberculosis to view all of the proteins identified in each sample, as well as to make direct comparisons manner, which gives information about the most abundant or most ionizable tryptic (MDR-TB) is a challenge to both clinicians and researchers. MDR-TB is defined as between each growth condition. Below is a screenshot of the Scaffold result file from the in vitro fragments in each sample. In order to get more quantitative information and to identify resistance to the two first-line drugs isoniazid (INH) and rifampin (RIF), and is linked experiment; highlighted are the statistically significant differentially expressed proteins found less prominent proteins (which may be obscured by more abundant ions), we will be to prolonged drug treatment, increased chance of surgery and a high rate of upon exposure of in vitro grown Mtb to drug. This includes the acyl carrier protein, AcpM, using MRM. This technique is highly specific, allowing us to monitor and collect data for mortality. In order to effectively develop new drugs, it is important to fully targeted peptides from the protein of interest. understand the mechanism of action of the current drugs, the methods by which which has the best ANOVA score. resistance develops, and the overall response within the bacterium. The majority of evidence related to changes in protein expression in response to drug treatment is Spectral counting is a method of label-free relative quantitation in complex samples, such as WCL. In general, if a protein is more abundant, spectra count, peptide count and sequence MRM Design correlated from microarray data. While the fluctuation in mRNA levels is often Designing a MRM experiment is not trivial – one or more tryptic peptides per protein indicative of protein expression, studies have shown that it is not always an accurate coverage will all increase; thus, this data is normalized to determine the quantities of protein representation of the proteome. For this reason, we have taken a proteomic approach relative to the population (1, 2). must be chosen for monitoring based on various rules. Shown below is an MRM design to discerning changes within the bacteria upon drug exposure. Utilizing Mtb H37Rv, Each lane represents the data from three for the AcpM protein. AcpM peptide and fragment ion pairs (transitions) were chosen INH, RIF and the combination of INH/RIF, was introduced in GAS media and technical replicates of one biological from a combination of experimental and purified protein MS/MS data. samples were grown as (biological) triplicates. Each sample was analyzed in triplicate replicate on a linear ion trap mass spectrometer directly coupled with nanoscale reverse phase For optimal results, MRM peptide candidates must abide by several rules: chromatography (nLC-LTQ). The raw data files were searched against the TB protein • Size: 7-30 amino acids in length Total Proteins sequence database using both the Sequest and Mascot database search engines. The ANOVA resulting analysis files were compiled in Scaffold. Initial quantitation by spectral Identified test • Ion charge of 2+ counting was performed to identify the dominant proteins expressed under each • Peptides must be unmodified (ex: Met can be oxidized) condition, as well as to distinguish the individual versus combined effects of each drug. Absolute quantitation on specific proteins of interest, including members of the • Peptides which tend to digest incompletely should be avoided FAS-II pathway using multiple reaction monitoring (MRM) analysis on a triple • Peptide must be unique to the protein of interest quadrapole mass spectrometer is underway. In addition, an in vivo translation of the experiment – involving H37Rv infected guinea pigs, is being completed to ensure that Analysis of the MS/MS data revealed 3 AcpM peptides in the recombinant protein digest the in vitro model correlates with the infection model and that the drug effect is and 5 peptides in the WCL digest. After applying these rules to the 5 peptides found in maintained. the AcpM analysis – two ideal peptides remained (Table below). Experimental Design Protein Peptide Observed mass range Average mass Actual mass range Charge Ion 1 Ion 2 AcpM IPDEDLAGLR 549.38 – 550.32 549.85 1096.74 – 1098.63 2+ Y8 888.26 Y9 985.3 H37Rv In vitro growth conditions H37Rv In vivo growth conditions TVGDVVAYIQK 597.81 – 598.44 598.13 1193.61 -1194.86 2+ Y9 992.32 Y6 721.28 Above: Table summarizing the two AcpM Venn diagram showing the unique and peptides and their properties – including two x3 x3 x5 x5 transition ions chosen for monitoring. shared proteins in this analysis. Of the 47 No Drug (control) Rifampin No Drug (control) INH/RIF shared proteins, AcpM represents one of 0.125 µg/mL the differentially expressed proteins. To the right, the MS/MS spectra for the peptides: IPDEDLAGLR – from which transition ions (purple boxes) were identified x5 x5 x3 x3 Isoniazid Rifampin Biological samples in triplicate The FAS-II Pathway (Y8 & Y9) and TVGDVVAYIQK (Y6 & Y9). Isoniazid Isoniazid/ 0.5 µg/mL Rifampin AcpM Below: The preliminary MRM data for both Aerosol Begin Drug Harvest INH AcpM peptides #1 and #2, showing a resolved Infection Treatments Lungs INH/RIF peak eluting at approximately 46.5 and 48 No drug minutes respectively. • Centrifuge cells Day 0 Day14 Day 28 RIF Rv0636 • Delipidate cells • Sonicate to prepare WCL • Homogenize lungs • Collagenase/DNase treatment Peptide # 1 Peptide # 2 • Delipidate cells Y9 Transition Y9 Transition • Sonicate to prepare WCL The chart above compares AcpM spectral counts in each biological replicate, clearly showing that Above: Preliminary MRM data from upon treatment with RIF the amount of AcpM During long chain fatty acid synthesis, AcpM serves as a Rv0636 dehydrase protein – not seen Sample preparation slighty increases. The increase is much more dramatic after exposure to INH. Unexpectedly, carrier protein. It presents the growing acyl chain to the various FAS-II proteins – allowing the acyl chain to undergo Peptide # 1 Y8 Transition Peptide # 2 Y6 Transition during normal LTQ analysis. Peptide and transitions were optimized on recombinant protein (data not shown). after subjecting the bacteria to both drugs, as in keto-reduction, dehydration, enoyl-reduction and finally a standard anti-mycobacterial chemotherapy, the condersation which adds an additonal 2 carbons to the chain. 50 – 100 ng peptides/injection LC-MS/MS effect is synergistic – drastically increasing the These acyl chains serve as the precursors to mycolic acids – a All of the initial MS work was performed on an LTQ mass spectrometer, including global levels of AcpM. vital component to the mycobacterial cell wall. Digest with 3 technical replicates analysis of in vitro WCL and in vivo lung homogenates. Traditionally MRM experiments are 10 µg WCL trypsin completed on a triple quadrupole mass spectrometer (QQQ), our preliminary MRM runs AcpM, The FAS-II Pathway & In vivo Results were executed on a Waters Xevo TQ. The use of MRM on either instrument will yield a higher sensitivity than the LTQ – allowing us to check for the presence of lower The acyl carrier protein, AcpM, is a key component in the fatty acid synthesis, type II abundance proteins in all in vivo samples. Also, with the use of isotopically-labeled peptide Reverse phase separation on a pathway (FAS-II)(shown above, right). Genes involved in the FAS-II pathway are found standards, absolute quantitation can be achieved. Zorbax 300SB-C18 column primarily in three operons: 1 - the kas operon, which is comprised of five genes, including over a 90-minute linear acetonitrile gradient acpM and two -ketoacyl-ACP-synthases, kasA and kasB (3); 2 - The mabA operon, which Future Data-dependent acquisition on contains the the two genes mabA and inhA, which encode the oxoacyl-ACP- reductase and Similar to microarray analysis, mass spectrometry allows us to compare different bacterial Nanospray-LTQ enoyl-ACP-reductase respectively (4); 3 - The last and most recently characterized operon states – i.e. drug exposed vs untreated. In the future, we would like to use the MRM encodes the three dehydrase proteins, Rv0635-Rv0637 (5). There is extensive evidence technique to quantitate various proteins in the FAS-II pathway, elucidating the changes implicating the inhibitory effects of INH on the FAS-II pathway. Previous microarray and which occur on the protein level. Additionally, establishment of MRM protocols will Data Analysis real time PCR experiments have shown a similar upregulation of the kas operon in response allow us to perform quantitation on virtually any given protein set. Unlike mRNA needed to treatment with INH (3). This observation gives validation to our experiment because we for microarrays, protein is much more stable, making these types of experiments feasable. Data were able to reproduce this expected response at the protein level and also because the Mining Mascot protein KasA is thought to be a target of INH. While we did not detect KasA in vitro, its Raw LTQ Data regulation should follow that of AcpM as they are integrated together in this operon. Acknowledgments (MS/MS Spectra) I would like to thank the entire Dobos lab, as well as the Tonge lab at Stony Brook University for providing purified SEQUEST In preliminary in vivo studies, 6 H37Rv infected guinea pig lung recombinant FAS-II proteins and expression plasmids. homogenates were analyzed. The AcpM protein was found to This work was supported by NIH Contract DHH5266200400091c and Gates Foundation Contract # 42775. Statistical be upregulated in long term infection (90 days; data to the left). Analysis Variation in spectral counts is attributed to the complexity of References X! Tandem the protein samples – containing both bacteria and host (1) Liu H, Sadygov RG, Yates JR 3rd.; Anal Chem. 2004 Jul 15;76(14):4193-201. proteins. We hypothesize that in drug treated guinea pigs, (2) htp://proteomics.btk.fi/uploads/media/Talk_3__Introduction_Spectral_Counting_03.pd AcpM and other FAS-II pathway components will be (3) Betts JC, McLaren A, Lennon MG, Kelly FM, Lukey PT, Blakemore SJ, Duncan K.; Antimicrob Agents upregulated in earlier infection time points (15 and 30 day) as Chemother. 2003 Sep;47(9):2903-13. (4) Marrakchi H, Ducasse S, Labesse G, Montrozier H, Margeat E, Emorine L, Charpentier X, Daffé M, Quémard A.; Spectral counts for 6 different 90 seen in the in vitro experiments. In order to quantitatively Microbiology. 2002 Apr;148 (Pt 4):951-60. Data was searched against a concatenated Mtb + IPI Mouse database containing day infected guinea pig lung 222,498 sequence entries compare these samples, we have chosen to use a directed, (5) Sacco E, Covarrubias AS, O'Hare HM, Carroll P, Eynard N, Jones TA, Parish T, Daffé M, Bäckbro K, Quémard samples. quantitative mass spectrometry technique – MRM. A.; Proc Natl Acad Sci U S A. 2007 Sep 11;104(37):14628-33. Epub 2007 Sep 5.
Pages to are hidden for
"Distinguishing the effects of Isoniazid and Rifampin on"Please download to view full document