In contrast to DNA microarrays, production of protein microarrays is an immense technological challenge due to high complexity and diversity of proteins. In this paper we investigate three essential aspects of protein microarray fabrication based on the highly parallel and non-contact TopSpot technology: evaporation of probes during long lasting production times, optimization of protein immobilization and improvement of protein microarray reproducibility. Evaporation out of the printhead reservoirs was reduced to a minimum by sealing the reservoirs with gas permeable foils or PDMS frames. This led to dramatically lowered setup times through the possibility of long-term, ready-to-print storage of filled printheads. To optimize immobilization efficiency 128 printing buffers were tested by printing two different proteins onto seven different microarray slide types. This way we were able to reduce the CV of spot diameter on the microarray slide below 1.14%. To remarkably increase protein immobilization efficiency on microarray slides the commonly used EDC-NHS system (a laboratory method for immobilization of proteins) was miniaturized by using a new drop-in-drop printing technique. Additionally the very fast UV cross-linking was used to immobilize antibodies. The optimized system was used to produce antibody microarrays and with it microarray ELISA experiments were performed successfully.

We present a technology for the highly parallel dispensing of a multitude of reagents. It allows one to dispense up to 96 different reagents simultaneously in a fixed array, in a volume range of 100 pL up to several nL. The pitch of the dispensed droplets can be as small as 500 microm. All channels are fired simultaneously, giving an unprecedented throughput. The system was originally developed for the high-throughput fabrication of microarrays, but can easily be adopted for other applications such as highly parallel filling of nanotiterplates. Based on our standard configuration we achieved droplets with 125- micro m in-flight diameter (1.2 nL) with a CV of <1%.

We report about the correlation between satellite free droplet release and liquid viscosity in a highly parallel, pressure driven nanoliter dispenser. In extensive studies, we found that for liquids of different viscosities the duration of the pressure pulse is the predominant effect compared to pressure amplitude. This result is of essential importance when actuation parameters have to be adopted for different media like oligonucleotide, DNA or protein solutions as it is the case for the non-contact high throughput fabrication of microarrays (Ducree et al., 2000). Experiments with oligonucleotides as well as with different proteins showed ascertained carry-over and cross-contamination free printing of DNA and protein microarrays. With it a prime critical point of microarray production is solved, leading to high quality whilst high throughput microarray fabrication. For oligonucleotides printing, we found CVs to be better than 1% within one single dispensing channel and 1.5% within all 24 channels of a 24 channel printhead for each used printing buffer. By optimizing the protein printing buffer the CVs for protein printing were reduced to about 1% within all 24 channels. As a serious practical application test oligonucleotides microarrays were produced using our nanoliter dispenser system. With it a full DNA hybridization experiment was performed. Clear positive signals one hand and no signals in the negative controls on the other hand showed that our system is suited for microarray production.

For the first time we report about the correlation between satellite free droplet release and liquid viscosity in a highly parallel, pressure driven nanoliter dispenser. In extensive studies we found that for liquids of different viscosities the length of the pressure pulse is the predominant effect compared to pressure amplitude. This result is of essential importance when actuation parameters have to he adopted for different media like oligonucleotide, DNA or protein solutions as it is the case for the non-contact highthroughput fabrication of microarrays [I]. For each used printing buffer we found the CV to be better than 1% within one single dispensing channel and 1 3 % within all 24 channels at a pitch of 500 pm.

No abstract