The upstream primer needs to contain minimally (from 5´ to 3´) a T7 promoter (5´-TAATACGACTCACTATAGGG-3´), a spacer region of 5-10 bases between the first transcribed base and the A of the initiation codon (ATG), and sequence corresponding to the target gene being amplified. The third G upstream from the 3´-end of the T7 promoter corresponds to the first base transcribed by T7 RNA Polymerase (Cat.# P2075). The addition of five random bases immediately upstream of the T7 promoter binding site can, in some cases, improve protein yields. In these cases, it is presumed that the additional bases at the extreme 5´ terminus of the DNA aid in transcription initiation by the T7 RNA Polymerase. The presence of 5-10 bases upstream of the initiation codon provides "room" for the 43S pre-initiation complex to bind and scan for the AUG initiation codon. In addition, the initiation codon may be placed in the context of a Kozak consensus sequence (A/GCCATGG). The Kozak consensus sequence is based on sequences found around the initiation codon of highly expressed genes and is believed to increase protein yields over initiation codons found in a non-Kozak consensus sequence. Finally, the upstream primer must have sufficient target sequence at its 3´-end to anneal to the correct template DNA (usually ~18 bases). Therefore, the overall length of a necessary upstream primer may be in excess of 53 bases.
Unless the target sequence being amplified contains a termination codon (TAA, TGA or TAG), it is necessary to include one of these codon sequences in the downstream primer. This facilitates efficient termination and generation of protein product of the correct size as well as recycling of the ribosomal subunits for additional rounds of translation. As with the upstream primer, it must have sufficient target sequence to anneal correctly to the template DNA (~18 bases).
It is best for both primers to be of similar length and %GC content (if possible) in order that they will possess similar annealing temperatures. The easiest way to increase the size of the downstream primer is to use more target sequence at the 3´-end (e.g., 48 bases instead of 18). There are multiple ways to calculate theoretical annealing temperatures (e.g., melting temperature, Tm) of oligonucleotide primers.
One equation considers the concentration of monovalent cations, %GC content and length of the oligo:
Tm = 81.5°C + 16.6°C × (log10[[Na+]+[K+]]) + 0.41°C × (%GC) - 675/N
...where N = number of nucleotides in the primer. Note that PCR is typically conducted in monovalent cation concentration of ~50m.
For example, the Promega T7 Promoter Primer (Cat.# Q5021) (TAATACGACTCACTATAGGG) has 5-T's, 7-A's, 4-C's and 4-G's (20mer, 40% GC). Thus, its melting temperature using the equation above is:
81.5°C + 16.6°C x (log10[0.05]) + 0.41°C × (%GC) - 675/20 = 42.5°C
Once the Tm for each primer is known, 5°C below the lowest Tm is a good starting point for the annealing step in PCR. If you see multiple products, try increasing the Tm by 1°C increments until the nonspecific products disappear.