Reverse-Phase High-Performance Liquid Chromatography Analysis

Reverse-Phase High-Performance Liquid Chromatography Analysis of Biomolecules 1 Theory RP-HPLC, or Reverse-Phase High-Performance Liquid Chromatography is a type of chromatography the features a liquid mobile phase, higher resolving power than traditional liquid chromatography, and a specific combination of stationary- and mobile-phases that results in separations that are the opposite of a “traditional” HPLC. Chromatography, in general, utilizes the flow of a “mobile-phase” containing the analyte through a “stationary-phase” – depending on the intermolecular forces of attraction between the individual analytes and the stationary- and mobile-phases, the various components of the sample will exit the chromatographic “column” at different times[?] . A substance that experiences greater intermolecular forces of attraction to the stationary-phase will be slowed, and will exit the column later than a substance without such attraction. If the mobile-phase is a liquid (as opposed to, say, a gas), the process is considered “liquid chromatography”; HPLC improves upon this process by pumping a pressurized liquid mobile-phase through an adsorbant stationary-phase, yielding far greater separation and resolving power than regular gravitationally-run liquid chromatography. The labelling “reverse-phase” is far more arbitrary in origin – historically, chromatographic columns utilized a polar stationary-phase and a nonpolar mobile-phase; “reverse-phase chromatography”, by contrast, uses a nonpolar stationary-phase and a polar mobile-phase, resulting in an inverted separation from “normal-phase” chromatography: more polar analytes will tend to elute earlier due to decreased interaction with the now-nonpolar stationary-phase. Because of its versatility, ease-of-use, and superior separation, HPLC is one of the most commonly-used classes of chromatography (itself one of the most common analytical methods), along with other methods like gas chromatography and column chromatography. 2 Results and Discussion This section contains only tabulated results from the Appendix. Derivations can be found in Appendix A on page 5. Graphs can be found in Appendix B on page 12. Raw data 1 Figure 1: Methionine, a nonpolar amino acid and one of our peptides can be found in Appendix A.1 on page 5. 2.1 Results Due to experimental failures that are detailed further in the appendix, we were unable to absolutely and quantitatively determine the concentration of peptides in our unknown sample. However, we were able to determine the ratios of peptides in our sample relative to the concentration of our calibration samples: Peak 1: 4.13498 40.6316 + 0.051785 0.012312 2 = 1690. ppm (relative) Peak 2: 11.96411 15.3068 + −0.02443 0.03288 2 2.2 = −131.3 ppm (relative) (?!) Accuracy and Error Of course, our accuracy in this experiment was exceedingly low. Ignoring the fact that we are missing a crucial piece of data to turn our relative concentration values into absolute ones (due to a failure for the instrument to read our sample, no less), the data we do have conspires to give us abysmal fits that lead to nonsensical data. A brief examination of our calibration curves shows that our error bars are gigantic and our R2 values abysmal (even, in one case, producing a negative slope that directly caused the apparently negative concentration of peptide seen above). Because of the vast amounts of missing data points in our tables (every N/A was a point not found on the instrument), we are simply not able to calculate anything with any degree of confidence. 2 Figure 2: Bradykinin, a blood-vessel-dilating peptide that is comparatively polar 2.3 Discussion While we cannot make many quantitative statements about the success of our experiment, we can still state certain facts with a reasonable degree of certainty. We can, for instance, determine the identity of the first and second peaks eluted by looking at the structures of the two peptides analyzed. Methionine, the first peptide, is a rather small essential amino acid and is largely nonpolar in character due to the presence of a fairly lengthy alkyl chain. Compared to the much larger bradykinin, enriched with numerous amines, amides, imines, and aromatic groups, we can say with a reasonable amount of certainty that the first peak is that of bradykinin and the second is methionine. Because we are performing reversephase HPLC, our stationary phase is a rather waxy resin with long alkyl chains that give it significant nonpolar character, whereas our mobile phases are comparatively polar in nature. As a result, the more nonpolar methionine is more slowed by intermolecular interactions with the stationary phase, allowing the more polar bradykinin to elute first. Our data was analyzed using the HPLC’s built-in light source, an Agilent 1100/1200 Diode Array Detector that filters light from a tungsten-filament bulb through a coupling lens, a deuterium lamp, an achromatic lens, a holmium oxide filter, a support lens, a flow cell, and a spectro lens befor finally passing through a programmable slit and onto a grating which projects a final spectrum[3] . Because our sample only called for UV irradiation and not visible light (a detail noticeable in the ACQ.TXT data reports), only the deuterium lamp is used. Because of the data loss that plagued this entire experiment, we are unable to calculate the niceties of our isocratic separations, such as the number of theoretical plates. 3 For the exact same reasons, we are unable to determine the amount of peptide in our unknown sample. We were able to determine the relative concentrations (1690. ppm and -131.3 ppm) by using the relative response factor, but even this approach is unhelpful due to abysmal correlation coefficients (ranging from 0.65 to 0.05) and simply nonsensical answers (negative concentration?!). While it is possible that our poor outputs are the result of detection limits and exceeding the linear range of the instrument, it is unlikely that our samples deviated so significantly from reasonable values that virtually none of them were able to construct a decent fit. Changing the wavelength of the detector would change the response factors of our analytes; methionine, containing little to no π-conjugation, is likely not an exceptional absorber compared to bradykinin, which has large, extended πsystems that likely absorb electrons of UV/Vis light readily. This also explains the comparatively lower absorption values of methionine compared to bradykinin. Nevertheless, the many flaws that riddled this experiment were likely a result of poor experimental preparation and instrumental methods rather than shortcomings of the HPLC instrument itself. Because our group was one of the last groups to complete this lab, we were forced to make do with insufficient quantities of peptide starting material, and whilst coordinating our sample preparation with another group we inadvertently screwed up the preparation of several of our samples. We most likely failed to get all of our results out of the HPLC due to a combination of this experimental error and incorrect setup of the HPLC methods, causing several of our key analytical data points to be dropped. Worst of all, we were unable to recreate and rerun our samples because we had run out of lab time over the semester with which to attempt to fix our mistakes. HPLC is a useful analytical method, even over other comparable separation methods like electrophoresis or other forms of chromatography, because it combines the relative ease-of-use of many chromatographic methods with the high-precision data that many more qualitative chromatographic methods lack. 4 A Calculations A.1 Peak Resolution Data RA,B = 2 ∗ tret,A −tret,B WA +WB RA,B =peak resolution tret =retention time W =baseline width A.1.1 1000 ppm std. Trial Method Peak 1 (min) Peak 2 (min) Peak 1 width Peak 2 width Resolution 1 C 1.134 2.311 0.202 0.166 6.3967391304 2 C N/A N/A N/A N/A N/A 3 C N/A N/A N/A N/A N/A A.1.2 500 ppm std. Trial Method Peak 1 (min) Peak 2 (min) Peak 1 width Peak 2 width Resolution 1 A 2.191 6.44 0.182 0.185 23.1553133515 2 A 2.022 6.06 0.207 0.179 20.9222797927 3 A 2.076 6.156 0.172 0.176 23.4482758621 5 A.1.3 100 ppm std. Trial Method Peak 1 (min) Peak 2 (min) Peak 1 width Peak 2 width Resolution 1 A 2.195 6.386 0.169 0.184 23.7450424929 2 A 2.063 6.165 0.18 0.19 22.172972973 3 A 2.044 6.241 0.161 0.199 23.3166666667 4 B N/A 2.19 0.151 0.157 7.025974026 5 B N/A 2.011 0.172 0.177 3.0601719198 6 B N/A 2.125 0.148 0.161 4.2459546926 7 C 1.112 2.198 0.164 0.149 6.9392971246 8 C N/A 2.198 N/A 0.16 N/A 9 C N/A 2.341 N/A 0.187 N/A A.1.4 10 ppm std. Trial Method Peak 1 (min) Peak 2 (min) Peak 1 width Peak 2 width Resolution 1 A 1.158 2.202 0.21 0.188 5.2462311558 2 A 1.465 2.039 0.165 0.204 3.1111111111 3 A 1.457 2.062 0.189 0.212 3.0174563591 A.1.5 1 ppm std. Trial Method Peak 1 (min) Peak 2 (min) Peak 1 width Peak 2 width Resolution 1 A 1.176 2.478 0.17 0.173 7.5918367347 2 A 1.518 2.129 0.193 0.201 3.1015228426 3 A 1.448 2.127 0.21 0.171 3.5643044619 A.1.6 0.1 ppm std. Trial Method Peak 1 (min) Peak 2 (min) Peak 1 width Peak 2 width Resolution 1 A 1.168 2.204 0.217 0.181 5.2060301508 2 A 1.481 2.067 0.127 0.185 3.7564102564 3 A 1.452 2.1 0.184 0.186 3.5027027027 6 A.1.7 Unknown Solution Trial Method Peak 1 (min) Peak 2 (min) Peak 1 width Peak 2 width Resolution 1 A 1.176 2.2 0.175 0.204 5.4036939314 2 A N/A 2.076 N/A 0.172 N/A 3 A 1.59 2.072 0.167 0.183 2.7542857143 A.1.8 Bradykinin std. Trial Method Peak 1 (min) Peak 2 (min) Peak 1 width Peak 2 width Resolution N/A N/A N/A N/A N/A N/A N/A Bradykinin Error Log 443 41e0 53462317 Method Method started: 67 41e0 53462319 Method CP Macro CP Macro 21:51:21 04/09/14 0 21:51:21 04/09/14 0 21:51:21 04/09/14 500 1 Error Method started 67 41e0 5346234b Method 0 Data not available; Data Analysis not done d5 1da7 5346234b ALS 21:51:21 04/09/14 Analyzing rawdata 080-2001.D 70 41de 53462349 21:50:33 04/09/14 500 Missing Vial, no Run possible 70 41de 53462349 21:50:31 04/09/14 0 1 Missing vial (80) 67 41e0 53462349 Method line# 20 vial# 80 inj# 1 Instrument running sample Vial 80 fb3 1da7 53462349 ALS 0 21:51:23 04/09/14 0 Method completed 21:51:23 04/09/14 7 A.1.9 Methionine std. Trial Method Peak 1 (min) Peak 2 (min) Peak 1 width Peak 2 width Resolution N/A N/A N/A N/A N/A N/A N/A Methionine Error Log 443 41e0 534625d5 Method Method started: 67 41e0 534625d7 Method ALS Method CP Macro CP Macro 0 22:02:57 04/09/14 0 22:02:57 04/09/14 500 1 Error Method started 67 41e0 53462603 22:02:57 04/09/14 0 Data not available; Data Analysis not done d5 1da7 53462603 A.2 22:02:56 04/09/14 Analyzing rawdata 081-2201.D 70 41de 53462601 22:02:15 04/09/14 500 Missing Vial, no Run possible 70 41de 53462601 22:02:13 04/09/14 0 1 Missing vial (81) 67 41e0 53462601 Method line# 22 vial# 81 inj# 1 Instrument running sample Vial 81 fb3 1da7 53462600 ALS 0 22:02:59 04/09/14 0 Method completed 22:02:59 04/09/14 Bradykinin and Methionine Calibration These calibration curves could not be created, due to a lack of any data (whatsoever) from their respective samples1 . As a result, we are completely unable to create calibration curves for these analytes, as well as calculating LoD or quantitative determination of the unknown peptide mixture, due to the lack of a known standard to use as a reference. Nevertheless, we will attempt to answer as many of these questions as possible using 1 See error logs on pages 7 and 8. 8 relative values. Based on the error logs, it is likely that human error was responsible for this data loss, as the samples simply failed to run. This could be caused by an incorrect method or by physical misplacement of the samples into the HPLC chamber. Whatever the cause, however, there is no way for us to make up for this lost information (by interpolation or the like) because there is simply no data to interpolate from. A.3 fi = Response Factor Calculation Ai f Astd std Because of the limitations detailed earlier regarding the peptide standards, we are unable to calculate absolute response factor and must settle for relative response factor: fri = fi fstd = Ai Astd We can calculate this value by plotting analyte concentration against response (in the form of area or height) and taking the slope of the best-fit line. Unfortunately, in this instance “best-fit” may be a misnomer. “Mediocre-fit” is probably more accurate. 9 10 Based on these graphs, we receive the following relative response factors (averaging out over multiple injections; injections were plotted separately because the data points were clearly stratified): Peak fri,1 fri,2,3 fri,avg Peak 1 .012312 .051785 .03205 Peak 2 .03288 A.4 -.02443 .00844 Relative Unknown Concentrations Using the previously calculated relative response factors, it is possible to calculate “relative unknown concentrations” in terms of the calibration solutions we created for curvefitting. Because our “1000 ppm” standard does not actually contain 1000 ppm of both peptides, these “relative concentrations” are ratios to the actual concentration of the standards created, something we could determine quantitatively if our reference values had not been lost. Peak 1: 40.6316 4.13498 + 0.051785 0.012312 2 = 1690. ppm (relative) Peak 2: 11.96411 15.3068 + −0.02443 0.03288 2 = −131.3 ppm (relative) (?!) 11