1	Purification for Quality In Pharmaceutical and natural product applications
2	Customer profile A variety of audiences with significantly differing requirements. High throughput low quantities single compound isolation. Solid phase synthesis. High Throughput variable quantities all compound isolation. Fermentation extracts. Walk up user up to 500mg reaction quantities single compound isolation. Solution Phase. Multigram purification of Chiral intermediates
3	Possible technologies Flash Chromatography. Prep HPLC conventional. Prep HPLC optimized with UV. Prep HPLC optimized with MS. Prep HPLC LC MS combination. Parallel Prep HPLC, UV with on line MS. SFC Solid bound reagents On line processes
4	UV directed Fractionation Too many fractions No chromophore No confirmation of product identity Most common cause of failure Sample with no chromophore. Precipitation of sample in the flow path usually at the sample loader. A reaction mixture only soluble in hot DMSO is not going to stay in solution during the chromatographic process!
5	Mass directed Fractionation Sample mass injection limited to approx 50mg. Peak volumes can be excessive. Peak purities can be compromised. No discrimination with chiral centres. Not suitable for Natural products. No ionization means no collection. High throughput limitations with hardware.
6	Quote from ACS Symposium presentation The most common failure modes for prep LC-MS purification are (a) failure to attain adequate chromatographic separation (resulting in impure product) and (b) failure to employ an appropriate MS fractionation threshold (with the result of either losing the sample because the threshold to set too high or collecting indiscriminately because the threshold is set too low).
7	UV vs MS fractionation
8
9	Extract from ACS Symposium Prep LC-MS is a complex technique and even a minor system fault may have serious implications for the user - ranging from failure to provide adequate product purity to loss of the sample. Consequently, we have developed a rigorous protocol to validate the entire prep LC-MS purification process and employ this protocol on a regular basis.
10	Biotage’s Position Clearly there are compromises. Both UV and MS directed can be successful. UV significantly lower cost of entry and cost of ownership. Lower technological barrier to entry in many cases with UV. More reliable ultimately? UV provides better quality of result? UV caters for serendipity. Unknowns are collected. Is combining UV with MS characterization optimal solution?
11	What is Quality? The totality of features and characteristics of a product or service that bear on its ability to satisfy stated or implied needs. Not to be mistaken for "degree of excellence" or "fitness for use" which meet only part of the definition. ISO 8402
12	Stated or implied needs In absolute terms 100% recovery of required compound(s). 100% purity of recovered compound(s). One fraction pr sample. 100% instrument reliability. Absolutely no risk of loss of sample. Walk up walk away simplicity of operation
13	Combinatorial Chemistry High degree of Automation. MTP format. Ability to interface with external devices. Reliability of operation. High % recovery of desired compound(s). Need to isolate and collect desired compound at high purity.
14	Medicinal Chemistry Walk up Walk away / open Access desired. Don’t collect too many fractions one is sufficient! Absolute reliability of operation. Don’t lose my sample!
15	Biotage Product Evolution Development of parallel flow path with simultaneous operation from two to four channels. Use of sophisticated fractionation algorithms that offers maximum recovery of product and given appropriate column and chromatographic conditions maximum purity. Software upgraded with closed loop feedback coding to prevent loss of valuable samples. Choice of flow cell path length. Eliminate excessive numbers of fractions. Multiple and stacked injection capability for large sample masses. Addition of MS characterization to eliminate the major issue of where is my compound?
16	The Parallex Flex 1-4 flow-streams (field upgradeable) Pumps 1-50ml/min (4000psi) Binary gradient (3^rd solvent 0 or 100%) Supports 1,2 and 3cm ID columns Independent methods/solvents
17	Robust Sample Injection Sample loader accepts deep well plates and vials Sandwich solvent plus air gap to minimize the risk of precipitation Wash option to eliminate contamination
18
19	Validation of Recoveries, Experimental Design Test loader for accuracy and consistency Weigh a microtiter plate filled with 1 mL of water (1 gram) per well Have Flex inject 1 mL Weigh plate and record data after each injection Repeat ~30 times. Data should be consistent and show 1 mL has been reproducibly injected. Test Flex for recovery using Customer sample (without column) Run at 5 mL/min Use Customer solvents, gradient, and decision criteria Repeat five times Dilute collected fractions to 10 mL Inject 20 µL into HPLC and compare to stock solution diluted to same volume. Recovery should be >90% Repeat process with modified decision criteria
20	Experimental Design cont’d C. Test Flex for recovery using Customer sample with column Use Customer solvents, gradient, and decision criteria Repeat ten times, alternating with a blank (DMSO) injection (400 µL) Sample decision criteria Threshold = 0.2 Slope = 0.2 Slope enable = 0.5 Blank decision criteria Threshold = 0.01 Slope = 0.01 Slope enable = 0.025 Dilute collected fractions to 50 mL (100 mL for propranolol) Inject 20 µL into HPLC and compare to stock solution diluted to same volume. Repeat (1) with threshold = 0.1, slope = 0.1, and slope enable = 0.25 Repeat (1) but inject 800 µL at half concentration (repeat 4 times) Repeat (1) but inject 1600 µL at one quarter concentration (repeat 3 times) Repeat (1) but use 5 µL air gap instead of DMSO sandwich and a 5 mL loader wash between runs (repeat 4 times) Repeat (1) with 5 µL air gap and no wash (repeat 4 times)
21	Results – Loader Accuracy Loader tested for injection volume precision and accuracy Thirty-six 1.0 mL injections of water from a deep-well microtiter plate made Plate weighed before and after each injection Loader is both accurate and reproducible.
22	Results – Recovery sans Column System recovery tested without column Eliminates impact of column on sample loss Customer gradient method used Detection at 220 nm Three decision criteria evaluated Threshold = 0.2, Slope = 0.2, Slope enable = 0.5 Threshold = 0.1, Slope = 0.1, Slope enable = 0.25 Threshold = 0.05, Slope = 0.05, Slope enable = 0.125 Flow rate 5 mL/min Results show excellent compound recovery at all decision levels
23	Recovery Summary
24	Summary The standard Flex configuration exceeds Customer’s minimum sample recovery criteria Specification = > 85% (terfenadine) Actual = 90.1% Sample recovery is enhanced by: Decreasing decision parameters to: Threshold = 0.1 Slope = 0.1 Slope enable = 0.25 Injecting larger volumes of more dilute samples Sample carryover is significantly diminished by incorporating a 5 mL loader wash between injections
25	The Parallex Flex: Superior Fractionation UV detection on 2 wavelengths (254, 280) others available including 219nm. Collection on threshold and or slope on both wavelengths Time window to restrict unwanted fractionation Superior algorithms to detect unresolved components for greater purity
26	Intelligent Fractionation
27
28	Superior Detection/Collection
29	Fraction Collection Dedicated collectors for greater productivity Wide variety of plates and test-tube racks Volumes automatically adjusted when flow rates are changed All plates/tube racks are bar coded for easy tracking Preloaded trays can be set up for easy use in open access
30
31
32
33
34	The Compound Factory
35	Flex for High Throughput Purification Up to 300 samples in 10 hours Sample tracking and data management Easily interfaced to synthesis and characterization systems Pilot™ and HTOC™ simplify post-purification processing Reduced capital equipment and running costs
36
37
38	The 0.1 mm flow cell
39	Development Partners
40	Objectives To link Parallex™and Parallex Flex™ to a mass spectrometer to enable on-line characterization The system is controlled by the Parallex Flex PC The software is updated as fractions are confirmed by yes/no Spectral data for all samples is readily accessible
41	Parallel Purification Systems Parallex Flex™ One to four independent flow paths Walk-up environment for multiple users Ideal for combinatorial arrays, small focused libraries and gm scale purification of chiral compounds.
42	Waters ZQ Mass Spectrometer
43	Interfacing the HPLC with ZQ
44	ZQ™ MUX™
45	Accurate™ Splitter
46	Parallex MUX™
47	Selecting OpenLynx Browser
48	Negligible cross-talk on four channels
49	Consistent data on all streams
50	Real Samples
51	Recovery of test compound
52	Confirmation of fractions
53	Summary The Parallex™ and Parallex Flex™ were successfully combined with the Waters ZQ™ using the MUX™ interface All control was through a single PC Fractions were automatically confirmed Yes/No Mass spectral data was available via Parallex/Flex Explorer Flex can run on 1,2 3 or 4 flow-streams
54	Acknowledgements Christine Edwards, Jack Liu, Pat Coffey and Bob Albrecht (Biotage) Dan Brooke, Andy Organ, Dave Hunter, and Tom Smith (GSK) Andrew Brailsford, Brian Smith, and Phil Kilby (Waters)