Improved Liquid Handling for MALDI Plate Spotting

by Mary Cornett PhD, Innovadyne Technologies, Inc.
and Alex Herrault, Susan Zhuang, Biosearch Technologies, Inc.

Introduction

Matrix-Assisted Laser Desorption Ionization–Time of Flight mass spectrometry (MALDI-TOF MS) is one of the main contemporary analytical methods for identifying proteins. It also plays a significant role in identifying the presence and purity of primers and other manufactured oligonucleotides.

While electrospray technologies have been introduced into the protein and oligo identification fields, its throughput is still not comparable to most MALDI systems, which continue to be considered the gold standard for compound identification. Furthermore, the need for larger sample volumes in most electrospray systems means that MALDI spotting is more economical for higher throughput applications.

For MALDI-TOF mass spectrometry, compounds must be spotted on a flat, metal MALDI target plate. Positional accuracy and precise deposition of the spots is critical in order for the ionizing laser to hit the sample. Unfortunately, classical liquid handlers using contact methods are prone to failure due to plate and tip irregularities and poor positional accuracy (typically > 100 µm/axis).

Some of the challenges involved in MALDI dispensing include:

The high salt content and evaporation speed of MALDI matrix solutions
The need for accurate spot placement on MALDI plates such that the laser hits the target reliably
The need for no contamination between samples.

Innovadyne’s platforms reliably fulfill MALDI spotting requirements with precisely manufactured motion control (<20 µm return to spot accuracy) and high-precision non-contact solenoid dispensing. The aspirate/dispense capabilities of the platform enable rigorous wash routines to be performed between samples.

Methods and Materials

We were able to compare an 8-channel Nanodrop liquid handler (Fig. 1) from Innovadyne Technologies with a Biomek FX conventional liquid handler from Beckman-Coulter under actual manufacturing use at Biosearch Technologies, Inc. (Novato, CA).

Fig 1. Nanodrop II Liquid Handler In Q/A Lab at Biosearch

Fig. 2 Nanodrop II Dispensing to a 384-Spot MALDI Plate

This particular facility uses MALDI technology for quality assurance of large numbers of synthetic oligonucleotides manufactured for its customers. Identical manufacturing runs were performed with both the Nanodrop system and the Biomek FX under the conditions noted in Table 1 and Table 2. It should be noted that the Nanodrop instrument used in this study was an 8-channel device as opposed to the 96-channels of the Biomek FX.

(While spotting time of the oligos was longer on the Nanodrop the shorter analysis time of the improved spots resulted in a shorter elapsed time for the spotting and detection process. Another system difference is the use of disposable tips on the Biomek FX compared to the use of fixed tips on the Nanodrop.)

Table 1. Common Conditions
*Condition*
MALDI plate type	Bruker AnchorChip 600/384
Concentration of oligos	5 µM
Oligo supply plate	Nalge Nunc 96-well skirted (VWR 73520-696)
Total number of samples	292

Table 2. Other Experimental Conditions
	*Nanodrop*	*Biomek FX*
Instrument software	Nanobuilder v. 2.4	Sagian AAO for Biomek FX
Matrix supply reservoir	Standard Nanodrop trough	Nalge Nunc 96-well skirted (VWR 73520-696)
Matrix spot volume	100 nL	500 nL
Oligo spot volume	50 nL	500 nL
Tips	Standard 125 µm fixed	Biomek AP96 P20 disposable

Results

Spotting Quality

One of the great advantages of low volume, non-contact dispensing for MALDI applications is that the smaller spot sizes are more uniform in size and dry more uniformly on the plate. Furthermore, with the accurate dispensing and <20 µm return-to-spot accuracy for the Nanodrop platform, spots are reliably placed and formed on the MALDI plates. Figure 3 clearly shows the accuracy and morphological differences of the two liquid handlers.

Fig 3a. Full Plate View, Plate Dispensed by Nanodrop

Fig. 3b. Full Plate View, Plate Dispensed by Biomek FX

Fig 3c. Closeup View, Plate Dispensed by Nanodrop

Fig. 3d. Closeup View, Plate Dispensed by Biomek FX

Figure 3. Full and closeup photos of Biomek and Nanodrop plates. These MALDI spotted plates were dispensed using the previously stated conditions for the Biomek FX (b & d) and for the Nanodrop (a & c).

Particularly noteworthy are the presence of doughnut-shaped spots in the Biomek spotting. This phenomenon is often found in larger volume spots that dry non-uniformly. The result of this type of morphology is that even if the spot does fall directly into the center of the well, the laser may miss the sample because it has dried primarily at the edges of the spotted area. Due to the uncertain location of the Biomek FX drops, a five times greater number of laser desorption shots are required for detection in comparison with the Nanodrop as these doughnut features were not observed in the Nanodrop spotting.

Data Quality

While most oligo batch identification rates were 100% for the Nanodrop instrument, a particularly challenging batch was chosen for this evaluation. Because of the greater accuracy of the Nanodrop spotting, there was a 93% identification success rate during this manufacturing run as opposed to 87% for the same batch of oligos spotted using the Biomek system. As can be seen in the example spectra below, both the mass accuracy and signal to noise ratios tend to be better for the Nanodrop instrument than for the Biomek instrument.

Fig 4a. Oligo # 3 (12590) Spectrum from Nanodrop-Dispensed Plate

Fig 4b. Oligo # 3 (12590) Spectrum from Biomek FX-Dispensed Plate

Figure 4: Example spectra from oligo of m/z 12590. Spectrum A is for the Nanodrop instrument, and is at 12547.51. Spectrum B is for the Biomek instrument at 12534.33.

Time Savings

As the results below indicate, the overall MALDI spotting time for the Nanodrop was longer than for the FX due to the fact that it is an 8-channel system as opposed to a 96-channel system (a 96-channel version of the Nanodrop, the Platemaker, is available from Innovadyne, and would spot a 384-spot MALDI plate with matrix and oligo in a total of about 4 minutes). A Platemaker was not in use at Biosearch, but its total process time can be estimated from the Nanodrop times and the cycle time of the device, and would be approximately half of that for the Biomek.

Even with the 12-fold difference in number of tips between the Nanodrop and the Biomek, the Nanodrop required only 70% of the total process time of the Biomek, as the result of differences in the laser desorption step. As can be seen in the previous pictures, Nanodrop spots are much more reliably centered on the MALDI plate, and the bulk of the spotted material is more reliably located in the center of the drop. These features make it much more likely that the laser will hit upon the sample spot on its first try. The longer time for analysis with the Biomek instrument reflects the need for the Biflex system to sample in 5 different positions in order to be assured of hitting the spots, as opposed to the single position needed for the Nanodrop instrument.

Table 3. Spotting and Laser Desorption Times
	*Biomek FX*	*Nanodrop*	*Platemaker (estimated)*
Matrix spot time	3 minutes	2 minutes	2 minutes
Oligo spot time	6 minutes	68 minutes (10s wash/ 5 second rinse per 8 samples)	2 minutes
Laser desorption time	316 minutes (50 shots)	161 minutes (10 shots)	161 minutes
Total process time	325 minutes	231 minutes	165 minutes

Cost Savings

In order for the Biosearch personnel to obtain satisfactory results from the Biomek spotting runs, it was set to dispense 500 nL drops of matrix and 500 nL drops of oligo, whereas the Nanodrop was set to dispense 100 nL of matrix and 50 nL of oligo. The five-fold reduction in reagents results in lower overall manufacturing costs. Additional savings were achieved on the Nanodrop due to the use of fixed tips. For this case, 5 boxes of Beckman Coulter Biomek AP96 P20 Pipette tips, at a cost of $4.64/box, gives a savings of $23.20 for a single plate. It is more difficult to quantify the cost savings of the higher throughput and increased identification rate.

Conclusion

Spotting of oligos and matrix solutions are ideal applications for the Nanodrop platform. The Innovadyne architecture requires no flow of sample through moving parts and so is ideal for high salt-content matrix solution as well as for limiting sample cross-contamination. Its non-contact, low volume capabilities as well as superior spot placement ability allow for consistent, centered spotting that drastically reduces required detection time and therefore overall process time. Because of the excellent spot placement and morphology, higher hit rates and better spectra are produced using the Nanodrop liquid handler relative to those of classical handlers. This reduces the need for repeated runs and time consuming alternative methods for confirming sample identities. The resulting MALDI process is shown to be a faster, less expensive, more reliable alternative to using classical liquid handlers such as the Biomek FX.