This is because there is a relationship between p-value and fold change. With no fold change (log2FC = 0) you have no significant difference hence p-value of 1 which is zero after -10log. With higher FC you can get lower p-values (hence higher -log10 values).
You are plotting the fold change of differential expression versus the confidence in detecting this differential expression. (q-value) Genes which have a low fold change will generally have a higher q-score.